Discussion:
North African DNA in the Iberian peninsula
Add Reply
Paulo Ricardo Canedo
2020-02-16 01:32:25 UTC
Reply
Permalink
This newsgroup has debated Islamic, Berber and Arab gateways in the Iberian peninsula. Well, this is not a specific link but DNA has shown that Portuguese, like me, and Spaniards do have North African DNA. It decreases from west to east. Western Iberians, like me, have between 5 and 13 percent. Central Iberians have between 2.5 and 5 percent. Eastern Iberians have between 1 and 2.5 percent. The Basques, who are not only from the far east of the peninsula but also from a mountainous region, have only between 1 and 2.5 percent. Strangely, there's not a North-South difference, with Northern Portugal (where I live), Galicia, the Asturias and Leon being among the regions with highest North African DNA. See https://www.eupedia.com/europe/autosomal_maps_dodecad.shtml#African for a map. Some of this North African DNA may have come from non-Islamic sources, though. Dear followers of the newsgroup, what do you think of all this?
j***@gmail.com
2020-02-16 02:12:43 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
This newsgroup has debated Islamic, Berber and Arab gateways in the Iberian peninsula. Well, this is not a specific link but DNA has shown that Portuguese, like me, and Spaniards do have North African DNA. It decreases from west to east. Western Iberians, like me, have between 5 and 13 percent. Central Iberians have between 2.5 and 5 percent. Eastern Iberians have between 1 and 2.5 percent. The Basques, who are not only from the far east of the peninsula but also from a mountainous region, have only between 1 and 2.5 percent. Strangely, there's not a North-South difference, with Northern Portugal (where I live), Galicia, the Asturias and Leon being among the regions with highest North African DNA. See https://www.eupedia.com/europe/autosomal_maps_dodecad.shtml#African for a map. Some of this North African DNA may have come from non-Islamic sources, though. Dear followers of the newsgroup, what do you think of all this?
First, I would say that there is no such thing as "north african DNA", except to mean "relatives of people who currently reside in northern africa". In simpler terms all you are saying is that people in western iberia are, on average, more closely related to people across the straight of Gibraltar than people in eastern Iberia. This is not surprising.

In even more simple terms, you are claiming that people, on average, are more closely related to each other the closer geographically they reside to each other. This is also unsurprising.

Since "Islamic" is a culture and religion, not a "race" of people, it is not really logical to try and figure out if the "northern african dna" was "islamic" or something else.

--Joe Cook
Peter Stewart
2020-02-16 02:24:22 UTC
Reply
Permalink
Post by j***@gmail.com
Post by Paulo Ricardo Canedo
This newsgroup has debated Islamic, Berber and Arab gateways in the Iberian peninsula. Well, this is not a specific link but DNA has shown that Portuguese, like me, and Spaniards do have North African DNA. It decreases from west to east. Western Iberians, like me, have between 5 and 13 percent. Central Iberians have between 2.5 and 5 percent. Eastern Iberians have between 1 and 2.5 percent. The Basques, who are not only from the far east of the peninsula but also from a mountainous region, have only between 1 and 2.5 percent. Strangely, there's not a North-South difference, with Northern Portugal (where I live), Galicia, the Asturias and Leon being among the regions with highest North African DNA. See https://www.eupedia.com/europe/autosomal_maps_dodecad.shtml#African for a map. Some of this North African DNA may have come from non-Islamic sources, though. Dear followers of the newsgroup, what do you think of all this?
First, I would say that there is no such thing as "north african DNA", except to mean "relatives of people who currently reside in northern africa". In simpler terms all you are saying is that people in western iberia are, on average, more closely related to people across the straight of Gibraltar than people in eastern Iberia. This is not surprising.
In even more simple terms, you are claiming that people, on average, are more closely related to each other the closer geographically they reside to each other. This is also unsurprising.
Since "Islamic" is a culture and religion, not a "race" of people, it is not really logical to try and figure out if the "northern african dna" was "islamic" or something else.
I think the point Paulo is suggesting is that the drive of "northern
African DNA" into Iberia was historially to a large extent Islamic.

Also, to my reading he has not claimed "that people, on average, are
more closely related to each other the closer geographically they reside
to each other", since he specifically noted that "Strangely, there's not
a North-South difference ...". Last I checked, North and South were
geographic concepts.

Peter Stewart
taf
2020-02-16 03:38:42 UTC
Reply
Permalink
While the west-east grade is new, the presence of significant 'North African' DNA in Iberia has been known for a decade. However, this comes with two major caveats.

First, it can't be dated, and there is significant uncertainty whether it represents primarily the Berbers that accompanied or followed the Muslim invasion, or instead includes a significant contribution from two earlier periods, the Phoenician/Carthaginian settlement and significant cross-strait traffic during the Roman era.

Second, is the 'North African' DNA signal really North African? I am not up on the latest results for Iberia, but at least one model was that a single ancient population spanned the straits, but that in Iberia this signal was diluted by repeated subsequent influx while their signal still represents a more significant portion of modern North Africans.

The take-home - it is very difficult to correlate DNA signals with specific populations in a location with a history of repeated invasion and influx. A decade of extensive sampling was needed to definitively identify the genetic signal of the Indo-Europeans. It takes a fine-grained temporal and geographic analysis of ancient DNA from dateable burials before and after each population wave. I have not been following it of late (and haven't had time to read the new paper yet), but the best study I have read only used three time points separated by several thousands of years each (one anotu 15,000 ago, one 5000, and one Roman-era) which is insufficient to sort out all of the different migrations.

taf
Peter Stewart
2020-02-16 07:37:15 UTC
Reply
Permalink
Post by taf
While the west-east grade is new, the presence of significant 'North African' DNA in Iberia has been known for a decade. However, this comes with two major caveats.
First, it can't be dated, and there is significant uncertainty whether it represents primarily the Berbers that accompanied or followed the Muslim invasion, or instead includes a significant contribution from two earlier periods, the Phoenician/Carthaginian settlement and significant cross-strait traffic during the Roman era.
Second, is the 'North African' DNA signal really North African? I am not up on the latest results for Iberia, but at least one model was that a single ancient population spanned the straits, but that in Iberia this signal was diluted by repeated subsequent influx while their signal still represents a more significant portion of modern North Africans.
The take-home - it is very difficult to correlate DNA signals with specific populations in a location with a history of repeated invasion and influx. A decade of extensive sampling was needed to definitively identify the genetic signal of the Indo-Europeans. It takes a fine-grained temporal and geographic analysis of ancient DNA from dateable burials before and after each population wave. I have not been following it of late (and haven't had time to read the new paper yet), but the best study I have read only used three time points separated by several thousands of years each (one anotu 15,000 ago, one 5000, and one Roman-era) which is insufficient to sort out all of the different migrations.
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.

But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.

Peter Stewart
Peter Stewart
2020-02-16 07:45:49 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
While the west-east grade is new, the presence of significant 'North
African' DNA in Iberia has been known for a decade.  However, this
comes with two major caveats.
First, it can't be dated, and there is significant uncertainty whether
it represents primarily the Berbers that accompanied or followed the
Muslim invasion, or instead includes a significant contribution from
two earlier periods, the Phoenician/Carthaginian settlement and
significant cross-strait traffic during the Roman era.
Second, is the 'North African' DNA signal really North African? I am
not up on the latest results for Iberia, but at least one model was
that a single ancient population spanned the straits, but that in
Iberia this signal was diluted by repeated subsequent influx while
their signal still represents a more significant portion of modern
North Africans.
The take-home - it is very difficult to correlate DNA signals with
specific populations in a location with a history of repeated invasion
and influx.  A decade of extensive sampling was needed to definitively
identify the genetic signal of the Indo-Europeans.  It takes a
fine-grained temporal and geographic analysis of ancient DNA from
dateable burials before and after each population wave.  I have not
been following it of late (and haven't had time to read the new paper
yet), but the best study I have read only used three time points
separated by several thousands of years each (one anotu 15,000 ago,
one 5000, and one Roman-era) which is insufficient to sort out all of
the different migrations.
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
On second thoughts, this could never work out - there would have been a
constant stream of people new to any region in the Arabian peninsula,
due to the annual influx of pilgrims passing through and to slaves
brought in from (mostly) Africa.

Peter Stewart
taf
2020-02-16 12:30:56 UTC
Reply
Permalink
Post by Peter Stewart
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
I think (going from memory), the genetic profile of the Arabian population from which Islam arose is pretty well understood - it is not dissimilar from that of Palestine at the same period, but with less 'Mediterranean' (e.g. Greek, Roman) influences. However, it would be problematic to use this as an 'Islamic' signature in the diaspora, where conversion and slavery played significant driving forces in the expansion. Of the non-native Muslim Iberians, Arabs formed an small segregated over-class, but demographically much more numerous were North African Berbers and Balkan slave-soldiers, and more significant yet were the muwallad native converts. Even with DNA this would be very difficult to sort out, though you could get a before-after comparison with enough sampling of burials from the post-Roman, pre-Islam era to compare with the modern inhabitants.

taf
JBrand
2020-02-16 17:26:16 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
I think (going from memory), the genetic profile of the Arabian population from which Islam arose is pretty well understood - it is not dissimilar from that of Palestine at the same period, but with less 'Mediterranean' (e.g. Greek, Roman) influences. However, it would be problematic to use this as an 'Islamic' signature in the diaspora, where conversion and slavery played significant driving forces in the expansion. Of the non-native Muslim Iberians, Arabs formed an small segregated over-class, but demographically much more numerous were North African Berbers and Balkan slave-soldiers, and more significant yet were the muwallad native converts. Even with DNA this would be very difficult to sort out, though you could get a before-after comparison with enough sampling of burials from the post-Roman, pre-Islam era to compare with the modern inhabitants.
taf
This recent article is only vaguely connected (i.e., same general geographical area), but interesting ...

https://www.cnn.com/2020/02/13/world/ancient-humans-africa-intl-scli-scn/index.html
Andrew Lancaster
2020-02-16 19:11:20 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
I think (going from memory), the genetic profile of the Arabian population from which Islam arose is pretty well understood - it is not dissimilar from that of Palestine at the same period, but with less 'Mediterranean' (e.g. Greek, Roman) influences. However, it would be problematic to use this as an 'Islamic' signature in the diaspora, where conversion and slavery played significant driving forces in the expansion. Of the non-native Muslim Iberians, Arabs formed an small segregated over-class, but demographically much more numerous were North African Berbers and Balkan slave-soldiers, and more significant yet were the muwallad native converts. Even with DNA this would be very difficult to sort out, though you could get a before-after comparison with enough sampling of burials from the post-Roman, pre-Islam era to compare with the modern inhabitants.
taf
For those semi-legendary Arabic male lines, in theory Y DNA should help and I know there are projects of amateurs out there trying to prove stuff. Not sure they have found much.

A potential problem here is that research companies and academic labs are now very happily able to focus on autosomal, whereas Y DNA in my opinion still has things to teach us.

From memory, when it comes to Iberia, there were hard-to-explain signs of Eastern Mediterranean and North African haplogroups in the area in and around the Basque country. (Sorry, no reference name comes to mind.) You will be much better informed than me, but when this information turned up I remember doing a bit of reading and learning about how there might well have been movements of smaller units of peoples into such areas. -- IF , it happened in the middle ages.

On the other hand, it has to be said, DATING these movements of people is still really difficult using any DNA. So it could have happened earlier. Methods to date the last common Y DNA ancestor of two people have improved, it should be said, and yet very little is done with this outside of a few of the more hardcore amateur groups.
Peter Stewart
2020-02-16 21:43:29 UTC
Reply
Permalink
Post by Andrew Lancaster
Post by taf
Post by Peter Stewart
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
I think (going from memory), the genetic profile of the Arabian population from which Islam arose is pretty well understood - it is not dissimilar from that of Palestine at the same period, but with less 'Mediterranean' (e.g. Greek, Roman) influences. However, it would be problematic to use this as an 'Islamic' signature in the diaspora, where conversion and slavery played significant driving forces in the expansion. Of the non-native Muslim Iberians, Arabs formed an small segregated over-class, but demographically much more numerous were North African Berbers and Balkan slave-soldiers, and more significant yet were the muwallad native converts. Even with DNA this would be very difficult to sort out, though you could get a before-after comparison with enough sampling of burials from the post-Roman, pre-Islam era to compare with the modern inhabitants.
taf
For those semi-legendary Arabic male lines, in theory Y DNA should help and I know there are projects of amateurs out there trying to prove stuff. Not sure they have found much.
A potential problem here is that research companies and academic labs are now very happily able to focus on autosomal, whereas Y DNA in my opinion still has things to teach us.
From memory, when it comes to Iberia, there were hard-to-explain signs of Eastern Mediterranean and North African haplogroups in the area in and around the Basque country. (Sorry, no reference name comes to mind.) You will be much better informed than me, but when this information turned up I remember doing a bit of reading and learning about how there might well have been movements of smaller units of peoples into such areas. -- IF , it happened in the middle ages.
On the other hand, it has to be said, DATING these movements of people is still really difficult using any DNA. So it could have happened earlier. Methods to date the last common Y DNA ancestor of two people have improved, it should be said, and yet very little is done with this outside of a few of the more hardcore amateur groups.
Can someone tell us whether or not the traceable origins of Y-DNA and
mt-DNA are found to differ more across the population in polygamous
societies than in others?

Peter Stewart
taf
2020-02-16 22:50:24 UTC
Reply
Permalink
Post by Peter Stewart
Can someone tell us whether or not the traceable origins of Y-DNA and
mt-DNA are found to differ more across the population in polygamous
societies than in others?
I don't know for certain, but I would expect there to be much higher mtDNA diversity in such populations than Y. The effective male population size in polygynous societies is going to be smaller than the effective female population size. The 'extra' males are murdered, castrated, exiled or sent off in armies to die or capture more territory with more women, while additional females are acquired through warfare, slave trade or other one-sided arrangements (in your typical marriage negotiation, it would be the polygynous society that would accept a bride into their population, while if they are to provide a groom, that groom would be joining the other population).

All of this is going to give you more diversity in the female lines than in the male lines. Due to the vicissitudes of genetic inheritance in small populations, many more Y lines will become extinct than mtDNA, and you are likely to have a much more recent most-recent common ancestor of Y than mtDNA (or alternatively, a few small clusters at the end of very long branches). Also, introduction of 'foreign' haplotypes, which has a large effect expanding overall diversity, will be almost exclusively mtDNA and not Y.

That is the expected pattern - I am sure the data are out there to do the formal analysis, of someone hasn't published it already, but I have not seen it.

taf
Peter Stewart
2020-02-17 03:11:06 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
Can someone tell us whether or not the traceable origins of Y-DNA and
mt-DNA are found to differ more across the population in polygamous
societies than in others?
I don't know for certain, but I would expect there to be much higher mtDNA diversity in such populations than Y. The effective male population size in polygynous societies is going to be smaller than the effective female population size. The 'extra' males are murdered, castrated, exiled or sent off in armies to die or capture more territory with more women, while additional females are acquired through warfare, slave trade or other one-sided arrangements (in your typical marriage negotiation, it would be the polygynous society that would accept a bride into their population, while if they are to provide a groom, that groom would be joining the other population).
All of this is going to give you more diversity in the female lines than in the male lines. Due to the vicissitudes of genetic inheritance in small populations, many more Y lines will become extinct than mtDNA, and you are likely to have a much more recent most-recent common ancestor of Y than mtDNA (or alternatively, a few small clusters at the end of very long branches). Also, introduction of 'foreign' haplotypes, which has a large effect expanding overall diversity, will be almost exclusively mtDNA and not Y.
That is the expected pattern - I am sure the data are out there to do the formal analysis, of someone hasn't published it already, but I have not seen it.
This was what I assumed, though of course unlike your reply not an
educated surmise on my part.

I wonder if various selection pressures might balance out the difference
in some cases - for instance, in parts of Turkey and Syria, and
especially in one region of Jordan, the local populations are largely
fair-haired and blue-eyed: this is said to be due to a strong preference
for Circassian wives over the past three centuries since Russian
imperialism scattered people from their homeland in the Caucasus making
them more vulnerable to slavery. So possibly the mt-DNA would be less
varied in those places, and mutatis mutandis in others where there may
have been different targets for human trafficking.

Slavery was not legally abolished in Saudi Arabia until the 1960s - and
of course continues today under a slightly altered guise in the
appalling mistreatment of foreign workers there and in Gulf states. Some
cultures have sadly failed to develop the vital organs for forming and
expressing public opinion in favour of decency. Just as sadly these
organs even when once healthy can weaken, as we see happening so
grotesquely in the US lately. Perhaps the mt-DNA profile of citizens
there may take an unnatural course in future too, from preferring
immigrants who don't come from countries the president disfavours.

Peter Stewart
Andrew Lancaster
2020-02-17 19:07:21 UTC
Reply
Permalink
Post by Peter Stewart
Post by Andrew Lancaster
Post by taf
Post by Peter Stewart
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
I think (going from memory), the genetic profile of the Arabian population from which Islam arose is pretty well understood - it is not dissimilar from that of Palestine at the same period, but with less 'Mediterranean' (e.g. Greek, Roman) influences. However, it would be problematic to use this as an 'Islamic' signature in the diaspora, where conversion and slavery played significant driving forces in the expansion. Of the non-native Muslim Iberians, Arabs formed an small segregated over-class, but demographically much more numerous were North African Berbers and Balkan slave-soldiers, and more significant yet were the muwallad native converts. Even with DNA this would be very difficult to sort out, though you could get a before-after comparison with enough sampling of burials from the post-Roman, pre-Islam era to compare with the modern inhabitants.
taf
For those semi-legendary Arabic male lines, in theory Y DNA should help and I know there are projects of amateurs out there trying to prove stuff. Not sure they have found much.
A potential problem here is that research companies and academic labs are now very happily able to focus on autosomal, whereas Y DNA in my opinion still has things to teach us.
From memory, when it comes to Iberia, there were hard-to-explain signs of Eastern Mediterranean and North African haplogroups in the area in and around the Basque country. (Sorry, no reference name comes to mind.) You will be much better informed than me, but when this information turned up I remember doing a bit of reading and learning about how there might well have been movements of smaller units of peoples into such areas. -- IF , it happened in the middle ages.
On the other hand, it has to be said, DATING these movements of people is still really difficult using any DNA. So it could have happened earlier. Methods to date the last common Y DNA ancestor of two people have improved, it should be said, and yet very little is done with this outside of a few of the more hardcore amateur groups.
Can someone tell us whether or not the traceable origins of Y-DNA and
mt-DNA are found to differ more across the population in polygamous
societies than in others?
Peter Stewart
I bet someone has done a study, but at least I can say there is now a lot of experience which says mt DNA and Y DNA disperse very differently in populations. Y DNA seems historically have moved quite quickly, with new lines arriving in an area and then almost eliminating others.

This may well be something which only began after the mesolithic when technology and travel perhaps became more connected to how successful men were? But that is pure speculation.
taf
2020-02-16 22:25:51 UTC
Reply
Permalink
Post by Andrew Lancaster
For those semi-legendary Arabic male lines, in theory Y
DNA should help and I know there are projects of
amateurs out there trying to prove stuff.
. . .
Post by Andrew Lancaster
On the other hand, it has to be said, DATING these
movements of people is still really difficult using any
DNA. So it could have happened earlier. Methods to date
the last common Y DNA ancestor of two people have
improved, it should be said, and yet very little is done
with this outside of a few of the more hardcore amateur
groups.
Such molecular clock dating is always statistical rather than absolute, and the variability is extremely high, so it is always problematic without calibration - an ancient specimen of known age that can be used to 'root' the calculation.

Just as an example of how hard this can be to sort out, Thomas Jefferson had a 'Middle Eastern' Y-chromosome, but what exactly does that mean - it could have come from a Medieval Jewish ancestor, an Iberian Arab captive of a viking raid that ended up in England through the viking slave trade, a centurion who came with a Roman army, or even a pre-historic Phoenician trader.

~~~~
Andrew Lancaster
2020-02-17 19:04:02 UTC
Reply
Permalink
Post by taf
Post by Andrew Lancaster
For those semi-legendary Arabic male lines, in theory Y
DNA should help and I know there are projects of
amateurs out there trying to prove stuff.
. . .
Post by Andrew Lancaster
On the other hand, it has to be said, DATING these
movements of people is still really difficult using any
DNA. So it could have happened earlier. Methods to date
the last common Y DNA ancestor of two people have
improved, it should be said, and yet very little is done
with this outside of a few of the more hardcore amateur
groups.
Such molecular clock dating is always statistical rather than absolute, and the variability is extremely high, so it is always problematic without calibration - an ancient specimen of known age that can be used to 'root' the calculation.
Just as an example of how hard this can be to sort out, Thomas Jefferson had a 'Middle Eastern' Y-chromosome, but what exactly does that mean - it could have come from a Medieval Jewish ancestor, an Iberian Arab captive of a viking raid that ended up in England through the viking slave trade, a centurion who came with a Roman army, or even a pre-historic Phoenician trader.
~~~~
Yes but the hardcore folk I allured to are using Y DNA SNPs with a lot of real sequencing instead of just markers, because they are looking for new SNPs. That is now doable. I have not seen any academics do it. It is of course still a statistical calculation, but the spread is getting smaller as such tests become common.
taf
2020-02-20 17:33:06 UTC
Reply
Permalink
Post by Andrew Lancaster
Post by taf
Such molecular clock dating is always statistical rather than absolute,
and the variability is extremely high, so it is always problematic without
calibration - an ancient specimen of known age that can be used to 'root'
the calculation.
Yes but the hardcore folk I allured to are using Y DNA SNPs with a lot of
real sequencing instead of just markers, because they are looking for new
SNPs. That is now doable. I have not seen any academics do it. It is of
course still a statistical calculation, but the spread is getting smaller
as such tests become common.
You seem to be talking about work done in a genealogically-relevant time period, which is really a distinct animal, as it provides its own calibration in the form of the paper-trail pedigree that is not available to scientists studying ancient population movement.

taf
Andrew Lancaster
2020-02-20 22:21:07 UTC
Reply
Permalink
Post by taf
Post by Andrew Lancaster
Post by taf
Such molecular clock dating is always statistical rather than absolute,
and the variability is extremely high, so it is always problematic without
calibration - an ancient specimen of known age that can be used to 'root'
the calculation.
Yes but the hardcore folk I allured to are using Y DNA SNPs with a lot of
real sequencing instead of just markers, because they are looking for new
SNPs. That is now doable. I have not seen any academics do it. It is of
course still a statistical calculation, but the spread is getting smaller
as such tests become common.
You seem to be talking about work done in a genealogically-relevant time period, which is really a distinct animal, as it provides its own calibration in the form of the paper-trail pedigree that is not available to scientists studying ancient population movement.
Right, but it could be used more widely eventually. The trees being discovered are indeed given a calibration by the cases with pedigrees, which is not all of them, but once they are calibrated, Y DNA male-line trees based on sequencing can go back thousands of years with a very high certainty of the correct branching, and a reasonable confidence, much better than past methods, on generation number estimates between branches.

If such projects were as big as the Y DNA projects in the good old days, the genealogical impact would be hard to ignore I think, but the numbers are small so far, with the testing companies diverting attention to themselves and their preferred products.

(I suppose we could say there is a similar slow down in real progress now in all the new internet genealogy technologies? Probably this is a bigger "business cycle" thing in many industries, with innovation at the bottom of the curve right now. Lots of companies that were start ups 15 years ago have some proven profitable ideas, and are feeling that trying to get monopoly power, or to resist those who have it, is the most important thing for now.)
s***@mindspring.com
2020-02-18 23:51:02 UTC
Reply
Permalink
Post by Andrew Lancaster
A potential problem here is that research companies and academic labs are now very happily able to focus on autosomal, whereas Y DNA in my opinion still has things to teach us.
Its too bad that the tools for doing Y-DNA genealogical research are still so primitive (unless something has recently arrived on the scene that I didn't notice). The raw data available online comes from so many different sources (mostly family projects run largely by amateurs) that it is virtually impossible to run a decent search for matches (except for your own kit on FTDNA) that any useful conclusions that might be obtained from the data are greatly diluted. I would love to see a reasonable GEDmatch-style utility for Y-DNA. I also wouldn't mind seeing something like that for mt-DNA, even though mt-DNA has limited usefulness as a genealogical tool. (My own closest mt-DNA matches are "genetic distance" 2, hundreds, perhaps even thousands, of years beyond any documentable time-frame).

Stewart Baldwin
Andrew Lancaster
2020-02-19 20:11:16 UTC
Reply
Permalink
Post by s***@mindspring.com
Post by Andrew Lancaster
A potential problem here is that research companies and academic labs are now very happily able to focus on autosomal, whereas Y DNA in my opinion still has things to teach us.
Its too bad that the tools for doing Y-DNA genealogical research are still so primitive (unless something has recently arrived on the scene that I didn't notice). The raw data available online comes from so many different sources (mostly family projects run largely by amateurs) that it is virtually impossible to run a decent search for matches (except for your own kit on FTDNA) that any useful conclusions that might be obtained from the data are greatly diluted. I would love to see a reasonable GEDmatch-style utility for Y-DNA. I also wouldn't mind seeing something like that for mt-DNA, even though mt-DNA has limited usefulness as a genealogical tool. (My own closest mt-DNA matches are "genetic distance" 2, hundreds, perhaps even thousands, of years beyond any documentable time-frame).
Stewart Baldwin
The most serious efforts are being made away from the big companies:

https://www.yfull.com/
See https://isogg.org/wiki/YFull

https://www.fullgenomes.com/
See https://isogg.org/wiki/Full_Genomes_Corporation
Denis Beauregard
2020-02-20 01:10:01 UTC
Reply
Permalink
Post by s***@mindspring.com
Post by Andrew Lancaster
A potential problem here is that research companies and academic labs are now very happily able to focus on autosomal, whereas Y DNA in my opinion still has things to teach us.
Its too bad that the tools for doing Y-DNA genealogical research are still so primitive (unless something has recently arrived on the scene that I didn't notice). The raw data available online comes from so many different sources (mostly family projects run largely by amateurs) that it is virtually impossible to run a decent
search for matches (except for your own kit on FTDNA) that any useful conclusions that might be obtained from the data are greatly diluted. I would love to see a reasonable GEDmatch-style utility for Y-DNA. I also wouldn't mind seeing something like that for mt-DNA, even though mt-DNA has limited usefulness as a genealogical
tool. (My own closest mt-DNA matches are "genetic distance" 2, hundreds, perhaps even thousands, of years beyond any documentable time-frame).

FTDNA has over 10,000 DNA "projects". A project is a collection
of results with some theme (surnames, geographical or haplogroup).
Projects allow to more experimented users to explore these results
and eventually to group them.

In my own projects for example and with the help of other admins,
we built a catalogue of Y and MT DNA signatures for gateway ancestors,
typically Quebec, Acadia and to some extent other French colonies.
That way, we may find for example that my own Beauregard line, which
is Jarret in France, is not too far from a Page family (common
ancestor estimated about year 400) and some Armenian people (common
ancestor about year 800).

So, it would be possible for a Spanish or Iberian project, for
example, to compare some Iberian Y-DNA (or MT) with that of
North African people, providing they all made a Big Y test or the
equivalent (but there is no mean to import Big Y style data into
FTDNA so you have to either do your own analysis or export to
Yfull or Ytree sites).


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
s***@mindspring.com
2020-02-22 23:29:16 UTC
Reply
Permalink
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.

The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.

Stewart Baldwin
Denis Beauregard
2020-02-23 06:12:53 UTC
Reply
Permalink
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer
important clues on how to narrow the search. How do you find these matches (if they exist)?

It is not necessary to be in a project to find a match.

If you are in a project, you can select matches only inside
that project and you can get Y-12 matches with GD=1.
Post by s***@mindspring.com
Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking
among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.

But there is some improvement for Big Y matches. Without being in the
same project, you can see who you are matching (if close enough)
Post by s***@mindspring.com
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for
his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.

Indeed, you must be a match to see more. This is a matter of privacy.
And if you have no test access (as a sponsor or as a project admin),
all you can do is to be lucky to get a Family Finder match and then
contacting that person.

But from time to time, I receive a request from someone having no
match and then I can forward an email to the testee, providing the
testee is in one of my projects and also he accepted to share
publicly his results.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
Andrew Lancaster
2020-02-24 08:30:44 UTC
Reply
Permalink
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.

Also see the reply of Dennis.

By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Peter Stewart
2020-02-24 08:53:25 UTC
Reply
Permalink
Post by Andrew Lancaster
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.
Also see the reply of Dennis.
By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Can you given an example where someone has tested ancient remains and
you are aware that this has helped medieval genealogy?

Peter Stewart
Andrew Lancaster
2020-02-26 14:43:38 UTC
Reply
Permalink
Post by Peter Stewart
Post by Andrew Lancaster
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.
Also see the reply of Dennis.
By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Can you given an example where someone has tested ancient remains and
you are aware that this has helped medieval genealogy?
Peter Stewart
For the most part this remains only a theoretical possibility so far. I suppose we can say that the Richard III case gave some genealogical conclusions, although not what we might have hoped for.
Denis Beauregard
2020-02-26 17:14:22 UTC
Reply
Permalink
On Wed, 26 Feb 2020 06:43:38 -0800 (PST), Andrew Lancaster
Post by Andrew Lancaster
For the most part this remains only a theoretical possibility so far. I suppose we can say that the Richard III case gave some genealogical conclusions, although not what we might have hoped for.
At this time, we have the yDNA signature of 2 sons of Louis XIII,
thanks to 3 tests from descendants.

http://www.francogene.com/triangulation/TRI0106.php

This signature was from a scientific publication. I suppose there
are some more similar DNA signatures one can get for royalty or
old nobility. So there is at least some kind of validation that
Y or MT DNA may achieve.

For Louis XIII, it is indeed possible both sons had the same
father who was not Louis XIII, so if the descendant of a cousin of
Louis XIII is tested, then the Louis XIII paternity will be
considered as proven (while the actual father could be a cousin
so there are some limitations to DNA).

For the British crown, since there is no legal limitation to DNA
testing, I am surprised there is no similar signature available.

The list of celebrities of Eupedia is unfortunately not as reliable
as it should. In most cases, it is the DNA of a relative who was
tested, and not that of the celebrity.

https://www.eupedia.com/genetics/famous_y-dna_by_haplogroup.shtml



Actually, it is weard to find that most of these DNA results are
from FTDNA while Eupedia said to avoid it...

There is also that forum, but no source is given.

https://www.eupedia.com/forum/threads/25236-Haplogroups-of-European-kings-and-queens


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
taf
2020-02-26 19:43:30 UTC
Reply
Permalink
On Wed, 26 Feb 2020 06:43:38 -0800 (PST), Andrew Lancaster wrote in
Post by Andrew Lancaster
For the most part this remains only a theoretical possibility so far.
I suppose we can say that the Richard III case gave some genealogical
conclusions, although not what we might have hoped for.
At this time, we have the yDNA signature of 2 sons of Louis XIII,
thanks to 3 tests from descendants.
There are a lot of cases of testing modern descendants to implicitly reveal information about earlier generations, but Peter was asking for something different, cases where testing of ancient remains has helped medieval genealogy. It is noteworthy that the testing of (claimed) ancient remains in the Bourbon case returned a result inconsistent with this testing of Louis XIII's descendants, though the problem almost certainly rests with the misattribution of the remains.

The only case that comes to mind was a paper from about 15 years ago that tested burials in an Austrian church and concluded that the genealogy of the counts involved needed to be rewritten because the relationships of the burials did not match the pedigree. However, this was a very poor study - they started out not knowing for sure who was buried in which sepulchre, plus their DNA testing was a bit shaky, so they were comparing iffy identifications to iffy results and drawing a conclusion from the apparent mismatch.

taf
taf
2020-02-26 22:36:50 UTC
Reply
Permalink
Post by taf
There are a lot of cases of testing modern descendants to implicitly reveal
information about earlier generations, but Peter was asking for something
different, cases where testing of ancient remains has helped medieval
genealogy.
I should add that of course, the Richard III case provided a single datum, but what to make of that datum is up in the air because we lack something solid to compare it to (a long-branch descent through two illegitimacies is not 'something solid'). It did seemingly confirm his maternity, but that was never really in doubt so I wouldn't call that helpful.

Comparable but better is work done on the 13th (?) century Mongol burials that I remember throwing a monkey wrench into some of what we thought we knew. It is better than Richard III because they were able to test multiple identifiable individuals. (I am being intentionally vague here, as I only have vague memories of the paper, other than that it suggested that the so-called Genghis Khan Y-haplotype was not found in these members of his patrilineage). That being said, not what people usually think of as 'medieval' even though it is contemporaneous.

taf
Peter Stewart
2020-02-26 22:47:03 UTC
Reply
Permalink
Post by taf
Post by taf
There are a lot of cases of testing modern descendants to implicitly reveal
information about earlier generations, but Peter was asking for something
different, cases where testing of ancient remains has helped medieval
genealogy.
I should add that of course, the Richard III case provided a single datum, but what to make of that datum is up in the air because we lack something solid to compare it to (a long-branch descent through two illegitimacies is not 'something solid'). It did seemingly confirm his maternity, but that was never really in doubt so I wouldn't call that helpful.
Comparable but better is work done on the 13th (?) century Mongol burials that I remember throwing a monkey wrench into some of what we thought we knew. It is better than Richard III because they were able to test multiple identifiable individuals. (I am being intentionally vague here, as I only have vague memories of the paper, other than that it suggested that the so-called Genghis Khan Y-haplotype was not found in these members of his patrilineage). That being said, not what people usually think of as 'medieval' even though it is contemporaneous.
This reminds me of the (sometimes unintentionally amusing) comments made
by biologists about various bird species that had been considered to
mate for life, until the paternity of their offspring shows to be
diverse and the females are then seen as actually promiscuous.

Nature is laughing in its sleeve at human understanding, always.

Peter Stewart
taf
2020-02-27 04:25:24 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
Comparable but better is work done on the 13th (?) century Mongol burials
that I remember throwing a monkey wrench into some of what we thought we
knew. It is better than Richard III because they were able to test
multiple identifiable individuals. (I am being intentionally vague here,
as I only have vague memories of the paper, other than that it suggested
that the so-called Genghis Khan Y-haplotype was not found in these members
of his patrilineage). That being said, not what people usually think of
as 'medieval' even though it is contemporaneous.
This reminds me of the (sometimes unintentionally amusing) comments made
by biologists about various bird species that had been considered to
mate for life, until the paternity of their offspring shows to be
diverse and the females are then seen as actually promiscuous.
Analogous results were seen with one of the big cat species. The traditional view was that each dominant male exercised exclusive control over the reproduction of their harem, while the young males were exiled and out of luck until they became strong enough to take over a harem from its aging dominant male. Then they tested the DNA and found out that the young males were sneaking in and cuckolding the dominant male with surprising frequency.
Post by Peter Stewart
Nature is laughing in its sleeve at human understanding, always.
In the case of the Genghis Khan haplotype, it wasn't so much human understanding as human rash assumption and hubris. It did not come from descendant testing, but rather a combination of population genetics and the sociological principle expounded by M. Brooks in his seminal work, History of the World, Part I. After leering, and . . . more, at a noblewoman, King Louis turns to the camera and says to the audience, "It's good to be the king".

Basically, one Y haplotype was found in a surprisingly large proportion of Central Asian men. What would cause a particular Y to be so favored? Societal power of the lineage founders, and since the khans were the male-line descendants of Genghis, the high-frequency Y chromosome must have been his. More logical deduction than evidence-based scientific conclusion.

taf
Peter Stewart
2020-02-26 22:37:29 UTC
Reply
Permalink
Post by taf
On Wed, 26 Feb 2020 06:43:38 -0800 (PST), Andrew Lancaster wrote in
Post by Andrew Lancaster
For the most part this remains only a theoretical possibility so far.
I suppose we can say that the Richard III case gave some genealogical
conclusions, although not what we might have hoped for.
At this time, we have the yDNA signature of 2 sons of Louis XIII,
thanks to 3 tests from descendants.
There are a lot of cases of testing modern descendants to implicitly reveal information about earlier generations, but Peter was asking for something different, cases where testing of ancient remains has helped medieval genealogy. It is noteworthy that the testing of (claimed) ancient remains in the Bourbon case returned a result inconsistent with this testing of Louis XIII's descendants, though the problem almost certainly rests with the misattribution of the remains.
The only case that comes to mind was a paper from about 15 years ago that tested burials in an Austrian church and concluded that the genealogy of the counts involved needed to be rewritten because the relationships of the burials did not match the pedigree. However, this was a very poor study - they started out not knowing for sure who was buried in which sepulchre, plus their DNA testing was a bit shaky, so they were comparing iffy identifications to iffy results and drawing a conclusion from the apparent mismatch.
This is about the state of progress I was expecting in the field of
applying DNA analysis to medieval genealogy.

It is easy to get carried away by scientific advancements into supposing
that these will necessarily shed light into dark corners, whereas in
reality illumination there may produce nothing more than novelty value.

Medieval genealogy is a pursuit of kinship as this has existed
throughout history, mainly as cultural and legal aspects of biological
facts that were not understood as subject to scientific study or proof.
Discovering "non-paternity" events may be interesting, but without
knowing even what the mother believed about her child's paternity they
are probably never going to be especially enlightening.

Peter Stewart
Andrew Lancaster
2020-02-29 23:06:17 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
On Wed, 26 Feb 2020 06:43:38 -0800 (PST), Andrew Lancaster wrote in
Post by Andrew Lancaster
For the most part this remains only a theoretical possibility so far.
I suppose we can say that the Richard III case gave some genealogical
conclusions, although not what we might have hoped for.
At this time, we have the yDNA signature of 2 sons of Louis XIII,
thanks to 3 tests from descendants.
There are a lot of cases of testing modern descendants to implicitly reveal information about earlier generations, but Peter was asking for something different, cases where testing of ancient remains has helped medieval genealogy. It is noteworthy that the testing of (claimed) ancient remains in the Bourbon case returned a result inconsistent with this testing of Louis XIII's descendants, though the problem almost certainly rests with the misattribution of the remains.
The only case that comes to mind was a paper from about 15 years ago that tested burials in an Austrian church and concluded that the genealogy of the counts involved needed to be rewritten because the relationships of the burials did not match the pedigree. However, this was a very poor study - they started out not knowing for sure who was buried in which sepulchre, plus their DNA testing was a bit shaky, so they were comparing iffy identifications to iffy results and drawing a conclusion from the apparent mismatch.
This is about the state of progress I was expecting in the field of
applying DNA analysis to medieval genealogy.
It is easy to get carried away by scientific advancements into supposing
that these will necessarily shed light into dark corners, whereas in
reality illumination there may produce nothing more than novelty value.
Medieval genealogy is a pursuit of kinship as this has existed
throughout history, mainly as cultural and legal aspects of biological
facts that were not understood as subject to scientific study or proof.
Discovering "non-paternity" events may be interesting, but without
knowing even what the mother believed about her child's paternity they
are probably never going to be especially enlightening.
Peter Stewart
However it is worth saying that with the types of testing which are POSSIBLE now, much more conclusive and interesting results would be possible.

But for now the best tests are often not only by archaeologists digging up a tomb in the steppes, but genealogists working on the Mid Western cousins.

The problem of course is privacy concerns for modern monarchs, and concerns about digging up bodies in the national heritage sites of modern countries. (Something medieval people did all the time!)

By the way I don't suppose there need to be lots of non-paternity events. Everyone expected this would turn up in genetic genealogy but the actual results have not turned out to be as racey as those cats taf mentioned.

It could of course be that royal families were different.
Paulo Ricardo Canedo
2020-03-01 01:07:59 UTC
Reply
Permalink
Post by Peter Stewart
Post by Andrew Lancaster
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.
Also see the reply of Dennis.
By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Can you given an example where someone has tested ancient remains and
you are aware that this has helped medieval genealogy?
Peter Stewart
Well, this is not a specific genealogy, but, in addition to the North African DNA in the Iberian peninsula, which is, at least, partially from the Islamic era, DNA has proven that Galicians, Northern Portuguese, Catalans and Valencians have more Germanic DNA than other Iberians. The Suebi and the Franks were more interested in settlement than the Visigoths.
Paulo Ricardo Canedo
2020-03-03 20:24:36 UTC
Reply
Permalink
Post by Peter Stewart
Post by Andrew Lancaster
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.
Also see the reply of Dennis.
By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Can you given an example where someone has tested ancient remains and
you are aware that this has helped medieval genealogy?
Peter Stewart
Well, this is also not a specific genealogy but it's based on ancient remains. Some East Germanic remains have shown relevant amounts of East Asian DNA. This fits with East Germanics' ties to the Huns. What do you think of it?
Peter Stewart
2020-03-03 22:05:26 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
Post by Peter Stewart
Post by Andrew Lancaster
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.
Also see the reply of Dennis.
By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Can you given an example where someone has tested ancient remains and
you are aware that this has helped medieval genealogy?
Peter Stewart
Well, this is also not a specific genealogy but it's based on ancient remains. Some East Germanic remains have shown relevant amounts of East Asian DNA. This fits with East Germanics' ties to the Huns. What do you think of it?
If you are asking me, I think it may be interesting for historical
demographers but not specific enough to be helpful for medieval
genealogists. What I was asking for is a problem that occurs in the
study of medieval genealogy, not glancingly incidental to this, and that
has been resolved or at least clarified by testing ancient remains as
suggested upthread. I can't imagine an example, and apparently
taf
2020-03-04 02:27:14 UTC
Reply
Permalink
Post by Peter Stewart
What I was asking for is a problem that occurs in the
study of medieval genealogy, not glancingly incidental to this, and that
has been resolved or at least clarified by testing ancient remains as
suggested upthread. I can't imagine an example, and apparently no-one
else here can either.
The main problem is that most burials from the period cannot be identified by name, and those that can are not allowed to be studied. There are a few cases of a medieval burial being identified by DNA analysis, like in the Richard III case or, depending on when you draw the cutoff, Copernicus.

It is hard enough to come up with a medieval person with a known burial around whom there is a standing genealogical question. Something like the maternity of William Longespee, if that wasn't learned a couple of decades ago through traditional documentation. Then you would need a to be allowed to do the exhumation, unlikely if the purpose was solely genealogical curiosity. And for most genealogical questions, you would need a second test subject as well. It doesn't surprise me there have been no instances.

taf
Peter Stewart
2020-03-04 04:58:21 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
What I was asking for is a problem that occurs in the
study of medieval genealogy, not glancingly incidental to this, and that
has been resolved or at least clarified by testing ancient remains as
suggested upthread. I can't imagine an example, and apparently no-one
else here can either.
The main problem is that most burials from the period cannot be identified by name, and those that can are not allowed to be studied. There are a few cases of a medieval burial being identified by DNA analysis, like in the Richard III case or, depending on when you draw the cutoff, Copernicus.
It is hard enough to come up with a medieval person with a known burial around whom there is a standing genealogical question. Something like the maternity of William Longespee, if that wasn't learned a couple of decades ago through traditional documentation. Then you would need a to be allowed to do the exhumation, unlikely if the purpose was solely genealogical curiosity. And for most genealogical questions, you would need a second test subject as well. It doesn't surprise me there have been no instances.
It doesn't surprise me at all, whereas the suggestion that there are
such instances was surprising - hence my question that has not been
answered with any example.

The number of medieval burials that can be guaranteed as undisturbed, so
that the remains must be those - and only those - of the one person
nominated as being placed in the grave more than five centuries ago,
would be miniscule. The number of substantial questions in medieval
genealogy that could conceivably be answered or even progressed by
presently-available DNA analysis from remains in these rare tombs would
be at most a tiny fraction of this miniscule proportion.

Peter Stewart
Denis Beauregard
2020-03-04 17:05:16 UTC
Reply
Permalink
Post by taf
The main problem is that most burials from the period cannot be identified by name, and those that can are not allowed to be studied. There are a few cases of a medieval burial being identified by DNA analysis, like in the Richard III case or, depending on when you draw the cutoff, Copernicus.
It is hard enough to come up with a medieval person with a known burial around whom there is a standing genealogical question. Something like the maternity of William Longespee, if that wasn't learned a couple of decades ago through traditional documentation. Then you would need a to be allowed to do the exhumation, unlikely if
the purpose was solely genealogical curiosity. And for most genealogical questions, you would need a second test subject as well. It doesn't surprise me there have been no instances.

There are 3 kinds of DNA tests for genealogy: autosomal, Y and MT.

Autosomal is usually considered as limited to 6 generations backward,
i.e. cousins you may find should be 12 generations distant, roughly
6 up and 6 down. For someone born 1900, 6 generations at 50 years each
go back to 1600 which is a limit as to where you can find a link
between 2 living persons (if their DNA was sampled in year 2000).
Typically, it would be someone born around 1940 and generations of
30 years. So, from a medieval burial (or someone buried before 1600
for this discussion), you may get a tiny link between that person
and a living person but likely of the noise level, i.e. too small
to be meaningfull.

However, the link between 2 persons of the same time could be
estimated. Is that the son/father of the other for example. But
you will need 2 burials for that purpose. This happens in some
medieval or old burials where there is some comparison of 2
persons found at the same place, but not enough to identify them.

Y and MT are more interesting. In both case, you must have a
continuous lineage to the tested person. That lineage doesn't need
to be from that person as it can be from a sibling, cousin, etc.

With Y, if you can perform a Big Y test (DNA must be in very good
state), you can get an estimate along the generations line. For
example, if John buried in 1400 has a brother Sam who has 2
living male descendants from 2 sons and each performs a Big Y,
then you can estimate the number of generations between John
and Sam (Big Y 700 is presumed to find 1 SNP per 2 generations).

Without a Big Y, you have nonetheless some matching between
John (the buried man) and the descendants of Sam (the reference
DNA). Less accurate that the Big Y 700, it will say if they are
probably (or not) of the same family. There are already matches
known for descendants of a common ancestor living before 1600.

The MT DNA is similar except you will need a female lineage from
the mother of John. If Sam is a full sibling or Mary is the sister
of John, then a descendant in matrilinear lineage of either Sam
or Mary will do the job.

In all these cases, you must know who was buried and the DNA
tests will only confirm it is **likely** that person.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
taf
2020-03-04 19:41:35 UTC
Reply
Permalink
On Tue, 3 Mar 2020 18:27:14 -0800 (PST), taf
Post by taf
The main problem is that most burials from the period cannot be identified by name, and those that can are not allowed to be studied. There are a few cases of a medieval burial being identified by DNA analysis, like in the Richard III case or, depending on when you draw the cutoff, Copernicus.
It is hard enough to come up with a medieval person with a known burial around whom there is a standing genealogical question. Something like the maternity of William Longespee, if that wasn't learned a couple of decades ago through traditional documentation. Then you would need a to be allowed to do the exhumation, unlikely if
the purpose was solely genealogical curiosity. And for most genealogical questions, you would need a second test subject as well. It doesn't surprise me there have been no instances.
There are 3 kinds of DNA tests for genealogy: autosomal, Y and MT.
Yes, and they can answer a question only if the burial(s) can be identified and permission granted for exhumation, and also if one can identify a second appropriate test subject, (whether it be a second burial or a living uni-parental lineage descendant). For autosomal, it requires two identifiable burials, as if one wasn't hard enough, because it is completely impractical for comparing a living person with medieval people.

This was not a question of what can be done in a perfect world. It was whether it has been done in this world, and it hasn't as far as I am aware, except for the one Austrian case that was poorly executed and produced uninterpretable results (which didn't stop the authors from interpreting them anyhow). (I went back and looked at the Mongol paper, and I was misremembering it - it wasn't an specifically identifiable burial, just identifiable as a member of a particular status/class (that in the society implied specific maternal and paternal lineages, but not specific parents within those lineages).

taf

taf
Peter Stewart
2020-03-04 22:25:09 UTC
Reply
Permalink
Post by taf
Post by taf
The main problem is that most burials from the period cannot be identified by name, and those that can are not allowed to be studied. There are a few cases of a medieval burial being identified by DNA analysis, like in the Richard III case or, depending on when you draw the cutoff, Copernicus.
It is hard enough to come up with a medieval person with a known burial around whom there is a standing genealogical question. Something like the maternity of William Longespee, if that wasn't learned a couple of decades ago through traditional documentation. Then you would need a to be allowed to do the exhumation, unlikely if
the purpose was solely genealogical curiosity. And for most genealogical questions, you would need a second test subject as well. It doesn't surprise me there have been no instances.
There are 3 kinds of DNA tests for genealogy: autosomal, Y and MT.
Autosomal is usually considered as limited to 6 generations backward,
i.e. cousins you may find should be 12 generations distant, roughly
6 up and 6 down. For someone born 1900, 6 generations at 50 years each
go back to 1600 which is a limit as to where you can find a link
between 2 living persons (if their DNA was sampled in year 2000).
Typically, it would be someone born around 1940 and generations of
30 years. So, from a medieval burial (or someone buried before 1600
for this discussion), you may get a tiny link between that person
and a living person but likely of the noise level, i.e. too small
to be meaningfull.
So for the purposes of medieval genealogy, autosomal DNA analysis is
about as practicable as is cold fusion for generating electricity?
Post by taf
However, the link between 2 persons of the same time could be
estimated. Is that the son/father of the other for example.
But does current science ask and answer "Is that the son/father..." or
only "Could that be/Is that most likely to be the son/father"?

<snip>
Post by taf
Y and MT are more interesting. In both case, you must have a
continuous lineage to the tested person. That lineage doesn't need
to be from that person as it can be from a sibling, cousin, etc.
With Y, if you can perform a Big Y test (DNA must be in very good
state), you can get an estimate along the generations line. For
example, if John buried in 1400 has a brother Sam who has 2
living male descendants from 2 sons and each performs a Big Y,
then you can estimate the number of generations between John
and Sam (Big Y 700 is presumed to find 1 SNP per 2 generations).
Without a Big Y, you have nonetheless some matching between
John (the buried man) and the descendants of Sam (the reference
DNA). Less accurate that the Big Y 700, it will say if they are
probably (or not) of the same family. There are already matches
known for descendants of a common ancestor living before 1600.
The MT DNA is similar except you will need a female lineage from
the mother of John. If Sam is a full sibling or Mary is the sister
of John, then a descendant in matrilinear lineage of either Sam
or Mary will do the job.
In all these cases, you must know who was buried and the DNA
tests will only confirm it is **likely** that person.
And if knowing who was buried always meant knowing whose remains are
currently in the burial site, this would be somewhat less problematic -
just as knowing which breed of pig was flying overhead centuries ago
might be helpful in another current thread.

Is medieval genealogy the study of anonymous linkages across eons of
time, or of continuous documented lineages at or from that time? If the
former, then perhaps DNA analysis can in future be helpful. If the
latter, then only some fantastic and unforeseeable advance in the field
could render it remotely useful. Like cold fusion.

Peter Stewart
taf
2020-03-05 00:53:05 UTC
Reply
Permalink
Post by Peter Stewart
So for the purposes of medieval genealogy, autosomal DNA analysis is
about as practicable as is cold fusion for generating electricity?
Yes, and no. A comparison of moderns to medievals cannot produce useful results, for the reasons described here in the past: too little shared DNA over that spean and the same DNA will have been found in too many people at the time period of the medieval person to be assumed to have come from them, and to 'triangulate' a reconstructed ancestral genome, as has been suggested here in the past, would require both a prohibitive number of people tested (hundreds of millions) and a good bit of luck. This does not apply to comparing two medieval burials.
Post by Peter Stewart
Post by Denis Beauregard
However, the link between 2 persons of the same time could be
estimated. Is that the son/father of the other for example.
But does current science ask and answer "Is that the son/father..." or
only "Could that be/Is that most likely to be the son/father"?
This question is asked and answered, though with a twist. One can determine that two samples unambiguously have a parent-child relationship, but unlike with intact DNA from living individuals, the DNA is invariably too degraded to determine which is the parent and which the child. This can only be resolved if the child is a juvenile, or by also identifying the other parent or another child. The approach has been used to determine the precise relationships in a bronze age mass grave, identifying parents with children, uncles and grandparents in what appears to have been a mass execution. Whether they are one degree removed, two, three, though, can be determined unambiguously. As you get farther, one can still detect there is a relationship, but the precise degree cannot be known for certain (the same is true with modern testing).
Post by Peter Stewart
And if knowing who was buried always meant knowing whose remains are
currently in the burial site, this would be somewhat less problematic -
just as knowing which breed of pig was flying overhead centuries ago
might be helpful in another current thread.
This is just one of the reasons why DNA studies have to be carefully interpreted, why you don't conclude Richard III was illegitimate because his Y chromosome didn't match that of someone removed by more than 20 generations, on paper. Had they shared the same Y, then it would have been strong evidence Richard was legitimate (though you can never know for certain), but there are many possible explanations for a mismatch.
Post by Peter Stewart
Is medieval genealogy the study of anonymous linkages across eons of
time, or of continuous documented lineages at or from that time?
I always viewed genealogy in any period as specific people being related to other specific people, which is why I can say both that most Iberian people descend from medieval Muslims, but that actual genealogical lines from medieval Muslims are extremely rare - one conclusion is about sociology, genetics and statistics, the other genealogy.
Post by Peter Stewart
If the
former, then perhaps DNA analysis can in future be helpful. If the
latter, then only some fantastic and unforeseeable advance in the field
could render it remotely useful. Like cold fusion.
Short of the invention of a time machine, . . . .

We already can do the analysis that would need to be done. The limitations are not of a type subject to scientific breakthrough, mostly coming down to the biological fact that 50% of DNA is inherited from each parent in each generation, that genetic information from burials is irreparably lost because medieval DNA has already degraded, and that the historical record is insufficient for most medieval burials to be unambiguously identified.

taf
Peter Stewart
2020-03-05 02:06:47 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
So for the purposes of medieval genealogy, autosomal DNA analysis is
about as practicable as is cold fusion for generating electricity?
Yes, and no. A comparison of moderns to medievals cannot produce useful results, for the reasons described here in the past: too little shared DNA over that spean and the same DNA will have been found in too many people at the time period of the medieval person to be assumed to have come from them, and to 'triangulate' a reconstructed ancestral genome, as has been suggested here in the past, would require both a prohibitive number of people tested (hundreds of millions) and a good bit of luck. This does not apply to comparing two medieval burials.
Post by Peter Stewart
Post by Denis Beauregard
However, the link between 2 persons of the same time could be
estimated. Is that the son/father of the other for example.
But does current science ask and answer "Is that the son/father..." or
only "Could that be/Is that most likely to be the son/father"?
This question is asked and answered, though with a twist. One can determine that two samples unambiguously have a parent-child relationship, but unlike with intact DNA from living individuals, the DNA is invariably too degraded to determine which is the parent and which the child. This can only be resolved if the child is a juvenile, or by also identifying the other parent or another child. The approach has been used to determine the precise relationships in a bronze age mass grave, identifying parents with children, uncles and grandparents in what appears to have been a mass execution. Whether they are one degree removed, two, three, though, can be determined unambiguously. As you get farther, one can still detect there is a relationship, but the precise degree cannot be known for certain (the same is true with modern testing).
Post by Peter Stewart
And if knowing who was buried always meant knowing whose remains are
currently in the burial site, this would be somewhat less problematic -
just as knowing which breed of pig was flying overhead centuries ago
might be helpful in another current thread.
This is just one of the reasons why DNA studies have to be carefully interpreted, why you don't conclude Richard III was illegitimate because his Y chromosome didn't match that of someone removed by more than 20 generations, on paper. Had they shared the same Y, then it would have been strong evidence Richard was legitimate (though you can never know for certain), but there are many possible explanations for a mismatch.
This is a crucial point that often gets overlooked - 20 generations
provide 20 opportunities for a non-paternity event separating the male
lines of a father and his legal son, and just one fewer opportunities
for another non-paternity event that coincidentally happened to reverse
this.
Post by taf
Post by Peter Stewart
Is medieval genealogy the study of anonymous linkages across eons of
time, or of continuous documented lineages at or from that time?
I always viewed genealogy in any period as specific people being related to other specific people, which is why I can say both that most Iberian people descend from medieval Muslims, but that actual genealogical lines from medieval Muslims are extremely rare - one conclusion is about sociology, genetics and statistics, the other genealogy.
This is exactly my view, which is why it is frustrating to ask my inital
question about genealogy only to have this prompt replies about
population studies.
Post by taf
Post by Peter Stewart
If the
former, then perhaps DNA analysis can in future be helpful. If the
latter, then only some fantastic and unforeseeable advance in the field
could render it remotely useful. Like cold fusion.
Short of the invention of a time machine, . . . .
We already can do the analysis that would need to be done. The limitations are not of a type subject to scientific breakthrough, mostly coming down to the biological fact that 50% of DNA is inherited from each parent in each generation, that genetic information from burials is irreparably lost because medieval DNA has already degraded, and that the historical record is insufficient for most medieval burials to be unambiguously identified.
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.

Peter Stewart
j***@gmail.com
2020-03-05 02:24:31 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
Post by Peter Stewart
So for the purposes of medieval genealogy, autosomal DNA analysis is
about as practicable as is cold fusion for generating electricity?
Yes, and no. A comparison of moderns to medievals cannot produce useful results, for the reasons described here in the past: too little shared DNA over that spean and the same DNA will have been found in too many people at the time period of the medieval person to be assumed to have come from them, and to 'triangulate' a reconstructed ancestral genome, as has been suggested here in the past, would require both a prohibitive number of people tested (hundreds of millions) and a good bit of luck. This does not apply to comparing two medieval burials.
Post by Peter Stewart
Post by Denis Beauregard
However, the link between 2 persons of the same time could be
estimated. Is that the son/father of the other for example.
But does current science ask and answer "Is that the son/father..." or
only "Could that be/Is that most likely to be the son/father"?
This question is asked and answered, though with a twist. One can determine that two samples unambiguously have a parent-child relationship, but unlike with intact DNA from living individuals, the DNA is invariably too degraded to determine which is the parent and which the child. This can only be resolved if the child is a juvenile, or by also identifying the other parent or another child. The approach has been used to determine the precise relationships in a bronze age mass grave, identifying parents with children, uncles and grandparents in what appears to have been a mass execution. Whether they are one degree removed, two, three, though, can be determined unambiguously. As you get farther, one can still detect there is a relationship, but the precise degree cannot be known for certain (the same is true with modern testing).
Post by Peter Stewart
And if knowing who was buried always meant knowing whose remains are
currently in the burial site, this would be somewhat less problematic -
just as knowing which breed of pig was flying overhead centuries ago
might be helpful in another current thread.
This is just one of the reasons why DNA studies have to be carefully interpreted, why you don't conclude Richard III was illegitimate because his Y chromosome didn't match that of someone removed by more than 20 generations, on paper. Had they shared the same Y, then it would have been strong evidence Richard was legitimate (though you can never know for certain), but there are many possible explanations for a mismatch.
This is a crucial point that often gets overlooked - 20 generations
provide 20 opportunities for a non-paternity event separating the male
lines of a father and his legal son, and just one fewer opportunities
for another non-paternity event that coincidentally happened to reverse
this.
My poor man's math says that studies of the distant past estimate an average rate of NPE at 1.5%. As roughly as possible the odds there was *not* any NPE over 20 generations at that rate would be 98.5% ^ 19 = 75%. So, even over 20 generations, the odds of the paper trail being correct if well documented, is probably at least better than even.

Another point I think was overlooked in some of these posts about comparing "ancient" DNA to a large database of modern people to try and 'find' descendants is that for some small scale test like Y-37, the confidence in any match is strongly cooborated by having the same surname or a backing paper trail. Independent of that it becomes extremely difficult to draw any conclusions at all testing so few samples about how many generations removed the two people may be.
--JC
taf
2020-03-05 03:19:19 UTC
Reply
Permalink
Post by j***@gmail.com
My poor man's math says that studies of the distant past estimate an average
rate of NPE at 1.5%. As roughly as possible the odds there was *not* any
NPE over 20 generations at that rate would be 98.5% ^ 19 = 75%. So, even
over 20 generations, the odds of the paper trail being correct if well
documented, is probably at least better than even.
Except the whole reason we are talking about this is that the Ys don't match - there _was_ an NPE (at least one) in those generations, the only question is whether it was in the four generations from Ed III to R III, or the 17 (or whatever) generations from Ed III to the living descendants. That turns your math on its head, it being more than 4 times as likely to have been in the long line as the short, just by generations.

As an aside, though I would also add that the NPE rate is likely not uniform across social groups, so it is a little bit tricky to generalize. In this case, the long branch includes two instances of a recognized illegitimate child. I think coming up with sufficient data to produce an NPE frequency among cases of recognized illegitimate children in the British upper class is a big ask. I don't even know if it would be higher or lower, I could see the conflicting societal influences and norms going either way, but I don't think it is safe to assume the NPE frequency would be the same as it would be for children born within a marriage.

taf
Denis Beauregard
2020-03-05 06:09:40 UTC
Reply
Permalink
Post by taf
As an aside, though I would also add that the NPE rate is likely not uniform across social groups, so it is a little bit tricky to generalize. In this case, the long branch includes two instances of a recognized illegitimate child. I think coming up with sufficient data to produce an NPE frequency among cases of recognized
illegitimate children in the British upper class is a big ask. I don't even know if it would be higher or lower, I could see the conflicting societal influences and norms going either way, but I don't think it is safe to assume the NPE frequency would be the same as it would be for children born within a marriage.

According to a study of Mr Larmuseau in Netherland or Belgium, the
frequency of NPE is higher in lower social classes. In a paper (sorry
I don't have the reference near me), he shown something like about
1% for richer people, and around 5-8% for poor people. In Quebec, we
have around 0.5% but this seems to be because many people who think
they have NPE would not join projects. Since more medieval genealogy
is about nobles and richer people, then they would have the best rate.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
taf
2020-03-05 02:53:57 UTC
Reply
Permalink
Post by Peter Stewart
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.
If you are talking about over a short number of generations, then autosomal does it just fine. If you are talking about medieval-to-modern, then no. The only complex analysis that would add power to the determination is what has been raised here, the possibility of reconstructing ancestral genomes by triangulation, but I did the calculations and you would have to test an impractical number of people to make it work well enough to give you enough reconstructed sequence to draw a meaningful conclusion. (but see below)

Three types of DNA have been mentioned, and the capabilities of each have been beat to death here at various times: autosomal, Y, and mt. There is a fourth that hasn't been mentioned much, and that is X. It is solely maternal for a male, but it acts like autosomal for a female, with half coming from each parent. This produces an unusual pattern, where a man's X DNA comes 50% from each maternal grandparent (none paternal), while a woman's comes 50% from her paternal grandmother and 25% each from each maternal grandparent. The proportions continue this disparate distribution as you go back. As a consequence, X chromosome DNA from a line that alternates male-female-male-female would be much more highly represented than any other line in the pedigree, while no line with two successive male generations would be represented at all.

If it just happened to be the precise gender-alternating line that you wanted to study, then though you start looking at only 1/20th of the amount of total DNA as you would studying autosomal, you only lose 50% every two generations, rather than every one with autosomal, and so at about 9 generations back, you pass the break-even point. Beyond that, you double your information over what you would have with autosomal with every generation back, but you are just about at the extreme of what autosomal can tell you, so maybe you get 12 or 13 or even 14 generations rather than 10 or 11, but that is about the extent of it. If at any point the line had two males in a row, using the X would be completely useless (like trying to trace the Y chromosome through a female), and any time you have two females in a row, you lose the 2-for-1 advantage of that generation, and given you are only getting a small number of extra generations as it is, that pretty much kills the advantage over autosomal. Maybe, though, if you combined this analysis with the reconstruction of ancestral genomes, you just might possibly be able to get there, but it would take an extraordinary amount of pedigree luck to have the gender-alternating line be the one you wanted, and also be able to identify enough other people similarly descended in gender-alternating lines branching at the right points to get the benefit. (This is all very much theoretical, with the calculation done on the fly.)

taf
s***@mindspring.com
2020-03-06 01:07:46 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.
If you are talking about over a short number of generations, then autosomal does it just fine. If you are talking about medieval-to-modern, then no. The only complex analysis that would add power to the determination is what has been raised here, the possibility of reconstructing ancestral genomes by triangulation, but I did the calculations and you would have to test an impractical number of people to make it work well enough to give you enough reconstructed sequence to draw a meaningful conclusion. (but see below)
Three types of DNA have been mentioned, and the capabilities of each have been beat to death here at various times: autosomal, Y, and mt. There is a fourth that hasn't been mentioned much, and that is X. It is solely maternal for a male, but it acts like autosomal for a female, with half coming from each parent. This produces an unusual pattern, where a man's X DNA comes 50% from each maternal grandparent (none paternal), while a woman's comes 50% from her paternal grandmother and 25% each from each maternal grandparent. The proportions continue this disparate distribution as you go back. As a consequence, X chromosome DNA from a line that alternates male-female-male-female would be much more highly represented than any other line in the pedigree, while no line with two successive male generations would be represented at all.
If it just happened to be the precise gender-alternating line that you wanted to study, then though you start looking at only 1/20th of the amount of total DNA as you would studying autosomal, you only lose 50% every two generations, rather than every one with autosomal, and so at about 9 generations back, you pass the break-even point. Beyond that, you double your information over what you would have with autosomal with every generation back, but you are just about at the extreme of what autosomal can tell you, so maybe you get 12 or 13 or even 14 generations rather than 10 or 11, but that is about the extent of it. If at any point the line had two males in a row, using the X would be completely useless (like trying to trace the Y chromosome through a female), and any time you have two females in a row, you lose the 2-for-1 advantage of that generation, and given you are only getting a small number of extra generations as it is, that pretty much kills the advantage over autosomal. Maybe, though, if you combined this analysis with the reconstruction of ancestral genomes, you just might possibly be able to get there, but it would take an extraordinary amount of pedigree luck to have the gender-alternating line be the one you wanted, and also be able to identify enough other people similarly descended in gender-alternating lines branching at the right points to get the benefit. (This is all very much theoretical, with the calculation done on the fly.)
When I was trying to understand my own X-chromosome matches (and those of my brother and sister), I found an alternate way of looking at the mathematics which indicates many similarities with the usual autosomal case. As you pointed out, in theory, a male gets all of his X-DNA from his mother, and a female gets half from each parent. So, setting aside the analysis for the male (which will essentially be the same as the analysis for his mother), let us look at the case for the female. She gets one half of her X-DNA from her mother, and the other half from her father (and therefore from her paternal grandmother). Thus, if we ignore males as being irrelevant as far as the X-chromosome is concerned (sorry guys!), a woman's two "X-parents" are her paternal grandmother and her mother (each of whom contributes half of her X-DNA), a woman's "X-grandparents" are her paternal grandmother's paternal grandmother, her paternal grandmother's mother, her mother's paternal grandmother, and her mother's mother, each of whom contributes one-fourth of her X-DNA, and so forth. Viewed in this way, X-DNA get inherited in a binary tree just as autosomal DNA does, except that some "X-generations" are one real generation (a female only) and some "X-generations" are two real generations (a male and his mother). So five "X-generations" back in the direct female line will be the same as five real generations, while five "X-generations" back in the strictly alternating male-female line will consist of ten real generations, each providing a theoretical 1/32 of the X-DNA, while each of the other 30 paths going five "X-generations" back (each providing a theoretical 1/32 of the X-DNA) will consist of between six and nine real generations. This explains, for example, why a 20 centimorgan X-chromosome match between two testees is, on average, less likely to correspond to a paper-trail match than a 20 centimorgan match on another chromosome, because the X-chromosome match is likely to be a larger number of real generations back, and therefore less likely to have a corresponding paper trail. This is an oversimplification which ignores some important factors (such as the differences between the individual chromosomes), but it did help me understand why it seemed that a smaller percentage of my long matches on the X-chromosome had a corresponding paper trail than what I expected from other chromosomes, and this approach helps to illustrate why there are some strong similarities between X-chromosome mathematics and standard autosomal mathematics.

Stewart Baldwin
P J Evans
2020-03-06 15:49:49 UTC
Reply
Permalink
Post by s***@mindspring.com
Post by taf
Post by Peter Stewart
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.
If you are talking about over a short number of generations, then autosomal does it just fine. If you are talking about medieval-to-modern, then no. The only complex analysis that would add power to the determination is what has been raised here, the possibility of reconstructing ancestral genomes by triangulation, but I did the calculations and you would have to test an impractical number of people to make it work well enough to give you enough reconstructed sequence to draw a meaningful conclusion. (but see below)
Three types of DNA have been mentioned, and the capabilities of each have been beat to death here at various times: autosomal, Y, and mt. There is a fourth that hasn't been mentioned much, and that is X. It is solely maternal for a male, but it acts like autosomal for a female, with half coming from each parent. This produces an unusual pattern, where a man's X DNA comes 50% from each maternal grandparent (none paternal), while a woman's comes 50% from her paternal grandmother and 25% each from each maternal grandparent. The proportions continue this disparate distribution as you go back. As a consequence, X chromosome DNA from a line that alternates male-female-male-female would be much more highly represented than any other line in the pedigree, while no line with two successive male generations would be represented at all.
If it just happened to be the precise gender-alternating line that you wanted to study, then though you start looking at only 1/20th of the amount of total DNA as you would studying autosomal, you only lose 50% every two generations, rather than every one with autosomal, and so at about 9 generations back, you pass the break-even point. Beyond that, you double your information over what you would have with autosomal with every generation back, but you are just about at the extreme of what autosomal can tell you, so maybe you get 12 or 13 or even 14 generations rather than 10 or 11, but that is about the extent of it. If at any point the line had two males in a row, using the X would be completely useless (like trying to trace the Y chromosome through a female), and any time you have two females in a row, you lose the 2-for-1 advantage of that generation, and given you are only getting a small number of extra generations as it is, that pretty much kills the advantage over autosomal. Maybe, though, if you combined this analysis with the reconstruction of ancestral genomes, you just might possibly be able to get there, but it would take an extraordinary amount of pedigree luck to have the gender-alternating line be the one you wanted, and also be able to identify enough other people similarly descended in gender-alternating lines branching at the right points to get the benefit. (This is all very much theoretical, with the calculation done on the fly.)
When I was trying to understand my own X-chromosome matches (and those of my brother and sister), I found an alternate way of looking at the mathematics which indicates many similarities with the usual autosomal case. As you pointed out, in theory, a male gets all of his X-DNA from his mother, and a female gets half from each parent. So, setting aside the analysis for the male (which will essentially be the same as the analysis for his mother), let us look at the case for the female. She gets one half of her X-DNA from her mother, and the other half from her father (and therefore from her paternal grandmother). Thus, if we ignore males as being irrelevant as far as the X-chromosome is concerned (sorry guys!), a woman's two "X-parents" are her paternal grandmother and her mother (each of whom contributes half of her X-DNA), a woman's "X-grandparents" are her paternal grandmother's paternal grandmother, her paternal grandmother's mother, her mother's paternal grandmother, and her mother's mother, each of whom contributes one-fourth of her X-DNA, and so forth. Viewed in this way, X-DNA get inherited in a binary tree just as autosomal DNA does, except that some "X-generations" are one real generation (a female only) and some "X-generations" are two real generations (a male and his mother). So five "X-generations" back in the direct female line will be the same as five real generations, while five "X-generations" back in the strictly alternating male-female line will consist of ten real generations, each providing a theoretical 1/32 of the X-DNA, while each of the other 30 paths going five "X-generations" back (each providing a theoretical 1/32 of the X-DNA) will consist of between six and nine real generations. This explains, for example, why a 20 centimorgan X-chromosome match between two testees is, on average, less likely to correspond to a paper-trail match than a 20 centimorgan match on another chromosome, because the X-chromosome match is likely to be a larger number of real generations back, and therefore less likely to have a corresponding paper trail. This is an oversimplification which ignores some important factors (such as the differences between the individual chromosomes), but it did help me understand why it seemed that a smaller percentage of my long matches on the X-chromosome had a corresponding paper trail than what I expected from other chromosomes, and this approach helps to illustrate why there are some strong similarities between X-chromosome mathematics and standard autosomal mathematics.
Stewart Baldwin
I got curious and built a 5-generation X-DNA chart, using ahnentafel numbers as names. Turns out you only need 8 generations. Reading in generations:
1
5,3
21,11,13,7
85,43,45,23,53,27,29,15
341,171,173,87,181,91,93,47,213,107,109,55,117,59,61,31
s***@mindspring.com
2020-03-06 15:57:28 UTC
Reply
Permalink
Post by P J Evans
Post by s***@mindspring.com
Post by taf
Post by Peter Stewart
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.
If you are talking about over a short number of generations, then autosomal does it just fine. If you are talking about medieval-to-modern, then no. The only complex analysis that would add power to the determination is what has been raised here, the possibility of reconstructing ancestral genomes by triangulation, but I did the calculations and you would have to test an impractical number of people to make it work well enough to give you enough reconstructed sequence to draw a meaningful conclusion. (but see below)
Three types of DNA have been mentioned, and the capabilities of each have been beat to death here at various times: autosomal, Y, and mt. There is a fourth that hasn't been mentioned much, and that is X. It is solely maternal for a male, but it acts like autosomal for a female, with half coming from each parent. This produces an unusual pattern, where a man's X DNA comes 50% from each maternal grandparent (none paternal), while a woman's comes 50% from her paternal grandmother and 25% each from each maternal grandparent. The proportions continue this disparate distribution as you go back. As a consequence, X chromosome DNA from a line that alternates male-female-male-female would be much more highly represented than any other line in the pedigree, while no line with two successive male generations would be represented at all.
If it just happened to be the precise gender-alternating line that you wanted to study, then though you start looking at only 1/20th of the amount of total DNA as you would studying autosomal, you only lose 50% every two generations, rather than every one with autosomal, and so at about 9 generations back, you pass the break-even point. Beyond that, you double your information over what you would have with autosomal with every generation back, but you are just about at the extreme of what autosomal can tell you, so maybe you get 12 or 13 or even 14 generations rather than 10 or 11, but that is about the extent of it. If at any point the line had two males in a row, using the X would be completely useless (like trying to trace the Y chromosome through a female), and any time you have two females in a row, you lose the 2-for-1 advantage of that generation, and given you are only getting a small number of extra generations as it is, that pretty much kills the advantage over autosomal. Maybe, though, if you combined this analysis with the reconstruction of ancestral genomes, you just might possibly be able to get there, but it would take an extraordinary amount of pedigree luck to have the gender-alternating line be the one you wanted, and also be able to identify enough other people similarly descended in gender-alternating lines branching at the right points to get the benefit. (This is all very much theoretical, with the calculation done on the fly.)
When I was trying to understand my own X-chromosome matches (and those of my brother and sister), I found an alternate way of looking at the mathematics which indicates many similarities with the usual autosomal case. As you pointed out, in theory, a male gets all of his X-DNA from his mother, and a female gets half from each parent. So, setting aside the analysis for the male (which will essentially be the same as the analysis for his mother), let us look at the case for the female. She gets one half of her X-DNA from her mother, and the other half from her father (and therefore from her paternal grandmother). Thus, if we ignore males as being irrelevant as far as the X-chromosome is concerned (sorry guys!), a woman's two "X-parents" are her paternal grandmother and her mother (each of whom contributes half of her X-DNA), a woman's "X-grandparents" are her paternal grandmother's paternal grandmother, her paternal grandmother's mother, her mother's paternal grandmother, and her mother's mother, each of whom contributes one-fourth of her X-DNA, and so forth. Viewed in this way, X-DNA get inherited in a binary tree just as autosomal DNA does, except that some "X-generations" are one real generation (a female only) and some "X-generations" are two real generations (a male and his mother). So five "X-generations" back in the direct female line will be the same as five real generations, while five "X-generations" back in the strictly alternating male-female line will consist of ten real generations, each providing a theoretical 1/32 of the X-DNA, while each of the other 30 paths going five "X-generations" back (each providing a theoretical 1/32 of the X-DNA) will consist of between six and nine real generations. This explains, for example, why a 20 centimorgan X-chromosome match between two testees is, on average, less likely to correspond to a paper-trail match than a 20 centimorgan match on another chromosome, because the X-chromosome match is likely to be a larger number of real generations back, and therefore less likely to have a corresponding paper trail. This is an oversimplification which ignores some important factors (such as the differences between the individual chromosomes), but it did help me understand why it seemed that a smaller percentage of my long matches on the X-chromosome had a corresponding paper trail than what I expected from other chromosomes, and this approach helps to illustrate why there are some strong similarities between X-chromosome mathematics and standard autosomal mathematics.
Stewart Baldwin
1
5,3
21,11,13,7
85,43,45,23,53,27,29,15
341,171,173,87,181,91,93,47,213,107,109,55,117,59,61,31
The way I was counting generations in my posting, this is four $X-generations" back from the original individual (16 individuals, each contributing 1/16 of the X-DNA).

Stewart Baldwin
Denis Beauregard
2020-03-05 06:20:45 UTC
Reply
Permalink
On Thu, 5 Mar 2020 13:06:47 +1100, Peter Stewart
Post by Peter Stewart
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.
One problem with autosomal is that DNA is not distributed uniformly
along the generations. You may get 16 to 33% from a grand-parent.
You can look at the chart at DNA Painter.

The number of positions you can scan is limited anyway. I don't think
you can get more than 6 generations back except if very lucky, for
example sharing a long segment and having only one lineage between
2 persons. Some people are reporting a relationship by 10 generations
back, but I think there are some re-combinations and there is more
than one lineage where the DNA arrived.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
Peter Stewart
2020-03-05 07:29:58 UTC
Reply
Permalink
Post by Denis Beauregard
On Thu, 5 Mar 2020 13:06:47 +1100, Peter Stewart
Post by Peter Stewart
Do you foresee any possibility of advances that may refine DNA analysis
to a point where this can allow the tracking of cross-gender linkages
through multiple generations? To me this seems like science fiction, but
I don't know enough to rule it out.
One problem with autosomal is that DNA is not distributed uniformly
along the generations. You may get 16 to 33% from a grand-parent.
You can look at the chart at DNA Painter.
The number of positions you can scan is limited anyway. I don't think
you can get more than 6 generations back except if very lucky, for
example sharing a long segment and having only one lineage between
2 persons. Some people are reporting a relationship by 10 generations
back, but I think there are some re-combinations and there is more
than one lineage where the DNA arrived.
Thanks - I didn't ask a question specific enough to learn what I was
after, so I will try again with two particular cases (below).

I understand that the answer with current technology is No, but what I
want to know is whether or not experts have confidence - or just
educated hope - that in the future this may conceivably be Yes.

1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?

2. If Charlemagne's remains were lacking, would the same question have a
different answer from analysing the DNA of a son of one of his daughters
and comparing this to Childeric's?

Peter Stewart
taf
2020-03-05 08:55:05 UTC
Reply
Permalink
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered. There is no way of parsing which of Charlemagne's DNA comes from his maternal grandfather without being able to test Charibert (and if we could do that it would render Charlemagne's DNA superfluous in asking whether Charibert descended from Childeric).

The only question we could ask using Charlemagne's is whether Charlemagne himself had some such descent. We are talking about somewhere in the neighborhood of 10 generations, and that is right on the edge of having a reasonable likelihood of detecting sufficiently-long segments to identify a relationship, were you using pristine modern genomic DNA, but the DNA in those burials is going to be damaged and that will reduce the ability to detect long enough shared segments to confirm a genealogical relationship. You would certainly find shared DNA between Childeric and Charlemagne, but it may not represent a direct genealogical connection. Any given stretch of shared DNA may just have been found at high probability across all of the 5th century Franks - one could potentially get a grip on this by digging up enough anonymous Frankish burials to get statistics on the general Frankish population.

It is also possible that one or the other (or both) won't have preserved DNA. Differences in temperature, humidity, etc. over the intervening millennium could cause the near-complete loss of sample DNA. There have been cases where they opened a tomb and found the entire body was a pile of dust, having been completely digested by contaminating bacteria, and you would get no DNA from that.

I don't imagine the situation improving via scientific breakthroughs - the problems don't arise from our inability to do the analysis, they have to do with the fundamentals of genetics and chemistry - you can't get around the nature of inheritance or the actions of chemistry (and bacteria) on DNA over a 1200/1500 year period. That being said, the reason they call them breakthroughs is that they weren't possible beforehand, and there have been a number of breakthroughs that have shown our entire understanding of what is possible to be wrong.
Post by Peter Stewart
2. If Charlemagne's remains were lacking, would the same question have a
different answer from analysing the DNA of a son of one of his daughters
and comparing this to Childeric's?
Two generations farther removed means you take whatever the slim chance of detecting a match would have been using Charlemagne and divide it by 4. Added to that, there is a 3/4 chance of any Merovingian DNA in the grandson coming from a grandparent other than Charlemagne. Just as with the Charlemagne/Charibert situation, you could only ask if that grandson descended from the Merovingians, not if Charlemagne did, by testing the grandson.

taf
Peter Stewart
2020-03-05 11:49:36 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered.
Thanks, this is exactly what I was wondering about - I assumed there was
no reasonable prospect of such a discovery, but of course I know a lot
less than you about what is now unknown much less what may conceivably
come to be known.

Some animal breeders base selective programs around alternating-gender
phenotypes, but evidently there is no traceable genetic underpinning to
this practice.

Peter Stewart
Denis Beauregard
2020-03-05 15:03:36 UTC
Reply
Permalink
On Thu, 5 Mar 2020 22:49:36 +1100, Peter Stewart
Post by Peter Stewart
Thanks, this is exactly what I was wondering about - I assumed there was
no reasonable prospect of such a discovery, but of course I know a lot
less than you about what is now unknown much less what may conceivably
come to be known.
To better understand what DNA can tell you, I would suggest to use
the DNA Painter tool at

https://dnapainter.com/tools/sharedcmv4

Enter 2000, 2100, 2200, etc. and see the predictions changing. With
a nearly complete DNA analysis with living people, you can't be sure
of their relation. With old DNA, one improvement is they can now
proceed with less losses, i.e. analyze longer segments. But that won't
change how DNA is transmitted and you can see from the DNA Painter
that the transmission is far from 50-50. So you may find that body 1
is father of body 2 who is fatehr of body 3, but with only bodies 1
and 3, you can't be sure 1 is grand-father of 3.

And as said by Taf, you can't say the relation is grand-father, uncle,
cousin, etc.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
Peter Stewart
2020-03-05 23:20:56 UTC
Reply
Permalink
Post by Denis Beauregard
On Thu, 5 Mar 2020 22:49:36 +1100, Peter Stewart
Post by Peter Stewart
Thanks, this is exactly what I was wondering about - I assumed there was
no reasonable prospect of such a discovery, but of course I know a lot
less than you about what is now unknown much less what may conceivably
come to be known.
To better understand what DNA can tell you, I would suggest to use
the DNA Painter tool at
https://dnapainter.com/tools/sharedcmv4
Enter 2000, 2100, 2200, etc. and see the predictions changing. With
a nearly complete DNA analysis with living people, you can't be sure
of their relation. With old DNA, one improvement is they can now
proceed with less losses, i.e. analyze longer segments. But that won't
change how DNA is transmitted and you can see from the DNA Painter
that the transmission is far from 50-50. So you may find that body 1
is father of body 2 who is fatehr of body 3, but with only bodies 1
and 3, you can't be sure 1 is grand-father of 3.
And as said by Taf, you can't say the relation is grand-father, uncle,
cousin, etc.
Thanks, but you are putting this in the present tense and the entire
point of my question was about what experts may consider to be future
possibilities.

Peter Stewart
Denis Beauregard
2020-03-06 01:47:49 UTC
Reply
Permalink
On Fri, 6 Mar 2020 10:20:56 +1100, Peter Stewart
Post by Peter Stewart
Thanks, but you are putting this in the present tense and the entire
point of my question was about what experts may consider to be future
possibilities.
Some improvements to come:

- faster and cheaper testing (more people will test)
- more data to compare (databases will grow)
- more complete testing with less DNA, i.e. it will be possible to
get usable DNA for genealogy from older and older remains. There
is probably a limit because bodies are not preserved eternally,
but some parts can survive longer if the conditions allow that
- separating each chromosome in a pair. In a typical autosomal test
from some labs, the values at each position are in alphabetical
order: if chromosome 1 has GGTAGC and the other of the same pair
has CCCCCC, then you will receive CG, CG, CT, AC, CG, CC. By
sampling much longer segments, it will be possible to separate
chromosomes. This means it will be possible to rebuild partially
the DNA of an ancestor. This may help to better identify the
relationship between 2 persons.
- longer segments for the Y chromosome. With Big Y500, there is
an average of 1 mutation (or SNP) per 4 generations. Big Y700
is estimated to give 1 SNP per 2 generations. Longer segments
would allow to have more STR and SNP to eventually identify
1 SNP per generation. That is, a possibility of identifying
where 2 male lineages are linked (common ancestor) before the
genealogical era, providing enough people are tested.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
Peter Stewart
2020-03-06 03:58:27 UTC
Reply
Permalink
Post by Denis Beauregard
On Fri, 6 Mar 2020 10:20:56 +1100, Peter Stewart
Post by Peter Stewart
Thanks, but you are putting this in the present tense and the entire
point of my question was about what experts may consider to be future
possibilities.
- faster and cheaper testing (more people will test)
- more data to compare (databases will grow)
- more complete testing with less DNA, i.e. it will be possible to
get usable DNA for genealogy from older and older remains. There
is probably a limit because bodies are not preserved eternally,
but some parts can survive longer if the conditions allow that
- separating each chromosome in a pair. In a typical autosomal test
from some labs, the values at each position are in alphabetical
order: if chromosome 1 has GGTAGC and the other of the same pair
has CCCCCC, then you will receive CG, CG, CT, AC, CG, CC. By
sampling much longer segments, it will be possible to separate
chromosomes. This means it will be possible to rebuild partially
the DNA of an ancestor. This may help to better identify the
relationship between 2 persons.
- longer segments for the Y chromosome. With Big Y500, there is
an average of 1 mutation (or SNP) per 4 generations. Big Y700
is estimated to give 1 SNP per 2 generations. Longer segments
would allow to have more STR and SNP to eventually identify
1 SNP per generation. That is, a possibility of identifying
where 2 male lineages are linked (common ancestor) before the
genealogical era, providing enough people are tested.
Thanks again, but I still haven't made myself clear - I am not asking
about refinements based on current knowledge and techniques (which I
have said repeatedly are understood not to allow for the answers I
enquired about in the cases I proposed), but rather about the likelihood
in the view of experts that new discoveries may yet come about and
change this.

Todd addressed the point directly when he wrote "absolutely not, not
now, not ever, unless an entirely unknown inheritance mechanism is
discovered".

In other words, in an expert opinion, there does not appear to be scope
in the currently unknown for this to happen. That is exactly what I
didn't know and wanted to know.

Peter Stewart
Andrew Lancaster
2020-03-06 12:31:26 UTC
Reply
Permalink
Post by Peter Stewart
Post by Denis Beauregard
On Fri, 6 Mar 2020 10:20:56 +1100, Peter Stewart
Post by Peter Stewart
Thanks, but you are putting this in the present tense and the entire
point of my question was about what experts may consider to be future
possibilities.
- faster and cheaper testing (more people will test)
- more data to compare (databases will grow)
- more complete testing with less DNA, i.e. it will be possible to
get usable DNA for genealogy from older and older remains. There
is probably a limit because bodies are not preserved eternally,
but some parts can survive longer if the conditions allow that
- separating each chromosome in a pair. In a typical autosomal test
from some labs, the values at each position are in alphabetical
order: if chromosome 1 has GGTAGC and the other of the same pair
has CCCCCC, then you will receive CG, CG, CT, AC, CG, CC. By
sampling much longer segments, it will be possible to separate
chromosomes. This means it will be possible to rebuild partially
the DNA of an ancestor. This may help to better identify the
relationship between 2 persons.
- longer segments for the Y chromosome. With Big Y500, there is
an average of 1 mutation (or SNP) per 4 generations. Big Y700
is estimated to give 1 SNP per 2 generations. Longer segments
would allow to have more STR and SNP to eventually identify
1 SNP per generation. That is, a possibility of identifying
where 2 male lineages are linked (common ancestor) before the
genealogical era, providing enough people are tested.
Thanks again, but I still haven't made myself clear - I am not asking
about refinements based on current knowledge and techniques (which I
have said repeatedly are understood not to allow for the answers I
enquired about in the cases I proposed), but rather about the likelihood
in the view of experts that new discoveries may yet come about and
change this.
Todd addressed the point directly when he wrote "absolutely not, not
now, not ever, unless an entirely unknown inheritance mechanism is
discovered".
In other words, in an expert opinion, there does not appear to be scope
in the currently unknown for this to happen. That is exactly what I
didn't know and wanted to know.
Peter Stewart
Hi Peter

According to my understanding, I agree with Todd's answer, but my answer was nevertheless possibly relevant to your question. My point was that there is one area (male lines) which already come to conclusions relevant for medieval genealogy, even without digging people up, but which can and will have stronger implications for medieval genealogy in the future.

Of course if we have to specifically prove a connection via someone's maternal side, as in your example, my comment is irrelevant.

Even now, if we had Childeric's remains, and Charlemagne's, we could tell if they were in the same male line.

Best Regards
Andrew
Andrew Lancaster
2020-03-05 20:06:48 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered.
Thanks, this is exactly what I was wondering about - I assumed there was
no reasonable prospect of such a discovery, but of course I know a lot
less than you about what is now unknown much less what may conceivably
come to be known.
Some animal breeders base selective programs around alternating-gender
phenotypes, but evidently there is no traceable genetic underpinning to
this practice.
Peter Stewart
If a really high resolution Y DNA test is done...

...and if we had a well identified ancient body...

...THEN we could come to conclusions about male line connections at least.

The highest resolution Y DNA tests possible (but not often used) will draw a male-line-only family tree with the correct branches, for any set of such tests, be they ancient of modern.

Generation numbers between the branches would however be estimates (but reasonable estimates).

Y DNA does not have the problem of autosomal DNA being mixed up every generation. You are then only dealing with a very small number of mutations each generation, distinguishing fathers and sons.
Peter Stewart
2020-03-05 23:21:52 UTC
Reply
Permalink
Post by Andrew Lancaster
Post by Peter Stewart
Post by taf
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered.
Thanks, this is exactly what I was wondering about - I assumed there was
no reasonable prospect of such a discovery, but of course I know a lot
less than you about what is now unknown much less what may conceivably
come to be known.
Some animal breeders base selective programs around alternating-gender
phenotypes, but evidently there is no traceable genetic underpinning to
this practice.
Peter Stewart
If a really high resolution Y DNA test is done...
...and if we had a well identified ancient body...
...THEN we could come to conclusions about male line connections at least.
The highest resolution Y DNA tests possible (but not often used) will draw a male-line-only family tree with the correct branches, for any set of such tests, be they ancient of modern.
Generation numbers between the branches would however be estimates (but reasonable estimates).
Y DNA does not have the problem of autosomal DNA being mixed up every generation. You are then only dealing with a very small number of mutations each generation, distinguishing fathers and sons.
Again, as in my last post, I was not asking about what can be done
today. My posts specifically stated this.

Peter Stewart
taf
2020-03-05 23:41:54 UTC
Reply
Permalink
Post by Peter Stewart
Some animal breeders base selective programs around alternating-gender
phenotypes, but evidently there is no traceable genetic underpinning to
this practice.
If I am following what you are talking about, this is probably an X chromosome effect on a recessive trait. For a male to express it they would need to get it from their mother, so to breed a herd of males expressing this, one would have to repeatedly pass the gene through females. I have already described how this plays out in a genetic genealogy context.

The other possibility for what you are describing is methylation. These are minor modifications to certain DNA at specific places, some of which depend on the gender of the parent from whom it was inherited.

<detailed description to follow, skip the next paragraph if you don't care for a fuller explanation>

If at a given site you had two options, R and W (for traits red vs white - I am not using standard genetics nomenclature here to make it easier to understand), and you were looking at a heterozygote with one of each, RW, for a gene unaffected by methylation it wouldn't matter if the R came from the father or the mother, you would get pink either way (red plus white). However if this was a material methylation-affected allele, then you would be either R*W, with the R coming from the mother turned off by methylation and you would display as white, or you would be RW*, with the W coming from the mother turned off, and you would present as red. (There are other ways the same basic pattern applies: the maternal methylation could turn on a gene that would otherwise be off, or the maternal gene could be the one with the methylation removed rather than added.) To see this in action, for real, compare a picture of a mule and a hinney to a horse and a donkey. Both mule and hinney are horse/donkey crosses, but one has a horse mother and the other has a donkey mother, and the differing methylation affects a whole host of characteristics, several of which (general body size, face length vs breadth, mane, leg length) are visibly different, more like one parent or the other in the two hybrids with identical DNA sequences (but not methylation).

Such parental-gender methylation differences, then, may be another reason livestock breeding would alternate generations. The thing is though, all of the parent-gender-specific methylation is scrubbed and reset at the time of sperm and egg formation, in each generation. Independent of the pattern of paternal and maternal methylation in a given individual, when they pass their DNA on, it will all have either the paternal or maternal methylation pattern, depending on their own sex. As a consequence, it is only theoretically possible that in the future a technique could be developed to identify which parent one got a particular piece of DNA from without testing the parent directly, but this would only work over a single generation, not a pedigree, so it is all but useless for genealogy.

taf
Peter Stewart
2020-03-06 04:06:51 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
Some animal breeders base selective programs around alternating-gender
phenotypes, but evidently there is no traceable genetic underpinning to
this practice.
If I am following what you are talking about, this is probably an X chromosome effect on a recessive trait. For a male to express it they would need to get it from their mother, so to breed a herd of males expressing this, one would have to repeatedly pass the gene through females. I have already described how this plays out in a genetic genealogy context.
The other possibility for what you are describing is methylation. These are minor modifications to certain DNA at specific places, some of which depend on the gender of the parent from whom it was inherited.
Thanks - I wasn't referring to scientific breeding programs but only to
intuitive ones.

Some breeders (fewer nowadays than in the past) select livetock to
retain based on physical characteristics that resemble their
opposite-gender parent.

The idea behind this was that such characteristics were more likely to
pass on predictably to their opposite-gender offspring, i.e. eventually
producing a peas-in-the-pod herd.

There are certainly examples where this result was achieved, but whether
because or in spite of the breeder's intentions, or by accident, or
somewhere in between, is another question. As is whether these breeders
were more or less likely to go broke than others ...

Peter Stewart
taf
2020-03-06 12:20:39 UTC
Reply
Permalink
Post by Peter Stewart
Thanks - I wasn't referring to scientific breeding programs but only to
intuitive ones.
Some breeders (fewer nowadays than in the past) select livetock to
retain based on physical characteristics that resemble their
opposite-gender parent.
Yes, I understood that. I just meant that the mechanism underlying their intuitive selection is probably either X-chromosome or methylation, even though the breeders would not have known that. The mule-hinney example is telling - for multiple millennia it was known that you crossed a female horse with a male donkey if you wanted a more horse-like hybrid, and vice versa. The patterns were evident to intuitive breeders, even though the reason underlying it, methylation, was only discovered in the past half-century.
Post by Peter Stewart
The idea behind this was that such characteristics were more likely to
pass on predictably to their opposite-gender offspring, i.e. eventually
producing a peas-in-the-pod herd.
There are certainly examples where this result was achieved, but whether
because or in spite of the breeder's intentions, or by accident, or
somewhere in between, is another question.
Yeah. This is pretty much the entire history of animal husbandry. Occasional outbreeding with massive inbreeding based on a combination of an understanding of basic inheritance and misunderstandings about complex genetic mechanisms, plus a good bit of random genetic drift, leading to 'fixation' of certain traits - the loss of all but one allele at each particular locus, so the population 'breeds true', and becomes a recognizable 'breed'.

taf
s***@mindspring.com
2020-03-06 15:52:33 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered.
Thanks, this is exactly what I was wondering about - I assumed there was
no reasonable prospect of such a discovery, but of course I know a lot
less than you about what is now unknown much less what may conceivably
come to be known.
If we relax the requirement "based on DNA only from these two burials" appearing in Todd's answer, there is an alternate scenario which, although improbable, cannot be ruled out. If we had usable DNA from both Charlemagne and Childeric, and Charlemagne really was a direct descendant of Childeric, and we had usable DNA from each of the intervening ancestors between them (admittedly a stretch, but not completely impossible), then the individual parent-child links would be verified by the DNA, and the genealogical link (in the biological sense) between Charlemagne and Childeric could be verified, along with the gender of each of the intervening links. The intervening remains would not have to be identified in advance as belonging to any specific individuals, and the extent to which the identity of the individual remains could then be verified would depend on the historical evidence. Even if remains from one or two of the intervening generations were absent, you might be able to make a reasonable circumstantial case.

To take a more general example, suppose that you had a large set of remains with usable DNA from the same general population. Assuming that there had not been excessive inbreeding in this population, any sets of full siblings that happened to be present among this population could be easily identified as full siblings from their DNA. Parent-child pairs could also be easily identified, although telling which was the parent and which was the child would be more difficult without additional information (any relevant full-sibling matches among the remains would settle the issue). More distant relationships could be inferred to some extent, but, with lower confidence about the exact relationship, with confidence about the exact relationships rapidly approaching zero as the number of generations with missing remains went up. So, if you were lucky enough that the group of remains came from enough members of an extended family, you could reconstruct some of the genealogical relationships. If some of the remains had already been identified with good evidence, you might be able to combine the reconstructed tree with historical evidence to identify some of the previously unidentified remains.

Going back to the Childeric-Charlemagne example, if taking DNA samples from a large number of remains happened to produce DNA from a full brother of Charlemagne (presumably Carloman), their mother (Bertha), and her father (Charibert/Heribert), then we would have a reasonable identification of Charibert's remains, and his Y-DNA would tell us whether or not he cam from the same male-line as Childeric. If we also happened to find DNA from Charibert's mother (another Bertha) and her unknown father (often conjectured to be a Merovingian), the latter's DNA would tell us whether or not he and Childeric were in the same male line. If they were, then that would be good circumstantial evidence (but not proof) that he was a Merovingian, but it would still not tell us the exact line of descent.

The biggest monkey wrench in the above scenario is that even if we included all known unidentified remains with usable DNA and assumed we had permission to test them all, we would not expect most family groups to be included with a large enough critical density to reconstruct the relationships within the groups, let alone identify them.

Stewart Baldwin
taf
2020-03-06 17:40:00 UTC
Reply
Permalink
Post by s***@mindspring.com
If we relax the requirement "based on DNA only from these two burials" appearing in Todd's answer, there is an alternate scenario which, although improbable, cannot be ruled out. If we had usable DNA from both Charlemagne and Childeric, and Charlemagne really was a direct descendant of Childeric, and we had usable DNA from each of the intervening ancestors between them (admittedly a stretch, but not completely impossible), then the individual parent-child links would be verified by the DNA, and the genealogical link (in the biological sense) between Charlemagne and Childeric could be verified, along with the gender of each of the intervening links.
Well, yes, and if my uncle had . . . never mind.

If you had the burials identified for each generation of a line connecting Charlemagne to Chideric, currently-available DNA analysis could prove it, but as I understood it, Peter's question was formulated in the way he did because in the real world we have identifiable tombs for Charlemagne and Childeric and perhaps nobody in between (if there even was an in between). It is not a minor change in this scenario to then conjure into existence all those unavailable generations of burials.

taf
s***@mindspring.com
2020-03-06 18:38:58 UTC
Reply
Permalink
Post by taf
Post by s***@mindspring.com
If we relax the requirement "based on DNA only from these two burials" appearing in Todd's answer, there is an alternate scenario which, although improbable, cannot be ruled out. If we had usable DNA from both Charlemagne and Childeric, and Charlemagne really was a direct descendant of Childeric, and we had usable DNA from each of the intervening ancestors between them (admittedly a stretch, but not completely impossible), then the individual parent-child links would be verified by the DNA, and the genealogical link (in the biological sense) between Charlemagne and Childeric could be verified, along with the gender of each of the intervening links.
Well, yes, and if my uncle had . . . never mind.
If you had the burials identified for each generation of a line connecting Charlemagne to Chideric, currently-available DNA analysis could prove it, but as I understood it, Peter's question was formulated in the way he did because in the real world we have identifiable tombs for Charlemagne and Childeric and perhaps nobody in between (if there even was an in between). It is not a minor change in this scenario to then conjure into existence all those unavailable generations of burials.
I agree that it is not a minor change. My main point was that we would not need the other remains to be identifiable in advance, as any random remains with a close enough relationship to either Childeric or Charlemagne would stand out from the test results alone. Obviously, this would require a great deal of luck. So, if we could arrange to test a large number of medieval corpses (extremely unlikely, but not impossible), mostly unidentified along with a few identified, it is a reasonable prospect that some of them would have identifiable relationships to each other. If an already identified body was included in such a match, this might lead to the identification of previously unidentified remains. The extent to which we could use this to bootstrap our way to the identification of a significant additional number is impossible to determine without actually doing it, but it most likely wouldn't amount to much. Still, I thought it was worth mentioning as a theoretical possibility.

Stewart Baldwin
Peter Stewart
2020-03-06 22:36:49 UTC
Reply
Permalink
Post by s***@mindspring.com
Post by taf
Post by s***@mindspring.com
If we relax the requirement "based on DNA only from these two burials" appearing in Todd's answer, there is an alternate scenario which, although improbable, cannot be ruled out. If we had usable DNA from both Charlemagne and Childeric, and Charlemagne really was a direct descendant of Childeric, and we had usable DNA from each of the intervening ancestors between them (admittedly a stretch, but not completely impossible), then the individual parent-child links would be verified by the DNA, and the genealogical link (in the biological sense) between Charlemagne and Childeric could be verified, along with the gender of each of the intervening links.
Well, yes, and if my uncle had . . . never mind.
If you had the burials identified for each generation of a line connecting Charlemagne to Chideric, currently-available DNA analysis could prove it, but as I understood it, Peter's question was formulated in the way he did because in the real world we have identifiable tombs for Charlemagne and Childeric and perhaps nobody in between (if there even was an in between). It is not a minor change in this scenario to then conjure into existence all those unavailable generations of burials.
I agree that it is not a minor change. My main point was that we would not need the other remains to be identifiable in advance, as any random remains with a close enough relationship to either Childeric or Charlemagne would stand out from the test results alone. Obviously, this would require a great deal of luck. So, if we could arrange to test a large number of medieval corpses (extremely unlikely, but not impossible), mostly unidentified along with a few identified, it is a reasonable prospect that some of them would have identifiable relationships to each other. If an already identified body was included in such a match, this might lead to the identification of previously unidentified remains. The extent to which we could use this to bootstrap our way to the identification of a significant additional number is impossible to determine without actually doing it, but it most likely wouldn't amount to much. Still, I thought it was worth mentioning as a theoretical possibility.
It is heartening that this newsgroup gets round to the discussion that
was hoped for by the enquirer.

My question was a bit like asking in the medieval era "How fast will
human beings ever be able to travel?" and hearing that this is limited
by the speed of horses.

There is a philosophical spin to my question along with the
bio-mechanical aspect: To what extent does nature cover its tracks?

Every individual who has ever lived (so far...) must have two biological
parents, and each progenitor back to the beginning of life must be
located in an exact position in the ancestry in order for that
individual to exist.

I am asking if there might ever be a possibility of determining that
position regardless of gender and without analysing DNA from all or even
most intervening generations.

From my understanding of genealogy, knowing that two individuals
separated by centuries derived their existence from the same all-male or
all-female lineage is such rudimentary information as to be practically
useless. Without filling in the intervening generations, this is not
genealogy but just demographic data and/or a shiny scientific novelty.

Maybe that is the limit of progress in this direction, and we will have
to plod along at maximum quadruped speed in this field for ever, relying
on documentary and circumstantial evidence for lack of useful scientif
taf
2020-03-06 23:09:36 UTC
Reply
Permalink
Post by Peter Stewart
My question was a bit like asking in the medieval era "How fast will
human beings ever be able to travel?" and hearing that this is limited
by the speed of horses.
For autosomal, I would suggest that it is more analogous to being limited by the speed of light. (Setting aside SB's scenario of finding enough unidentified burials to test that out of the large number of people during those 300 or so years so that one might just happen to be in the immediate line connecting Charlemagne to Childeric.) Unless the science is not what everyone agrees it is, it is a physical limitation (in one case physics, in the other the actual chemistry underlying genetics) that stands in the way, not a technical one or one that can be overcome by ingenuity. While one can in some cases distinguish which parent DNA came from, there is simply no distinguishing characteristic of DNA that lets you identify which DNA came from which grandparent without testing another person in the immediate proximity.

With the uniparental lines, one could produce a branched tree using genetic approaches and modern descendants, and compare it to a branched tree using paper-trail genealogy, and potentially line them up to identify the nodes of the molecular tree with the names in the paper tree, but for the medieval period available sampling and trustworthy trees are usually insufficient to be sure without digging up multiple people.

taf
Peter Stewart
2020-03-06 23:45:55 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
My question was a bit like asking in the medieval era "How fast will
human beings ever be able to travel?" and hearing that this is limited
by the speed of horses.
For autosomal, I would suggest that it is more analogous to being limited by the speed of light. (Setting aside SB's scenario of finding enough unidentified burials to test that out of the large number of people during those 300 or so years so that one might just happen to be in the immediate line connecting Charlemagne to Childeric.) Unless the science is not what everyone agrees it is, it is a physical limitation (in one case physics, in the other the actual chemistry underlying genetics) that stands in the way, not a technical one or one that can be overcome by ingenuity. While one can in some cases distinguish which parent DNA came from, there is simply no distinguishing characteristic of DNA that lets you identify which DNA came from which grandparent without testing another person in the immediate proximity.
With the uniparental lines, one could produce a branched tree using genetic approaches and modern descendants, and compare it to a branched tree using paper-trail genealogy, and potentially line them up to identify the nodes of the molecular tree with the names in the paper tree, but for the medieval period available sampling and trustworthy trees are usually insufficient to be sure without digging up multiple people.
Yes, but of course the physical limitations of travel were not what
everyone agreed these to be in the medieval era. Can today's science
really tell us that a point has been reached where the known outweighs
the unknown? This may be the case, as scientific advances tend to arrive
at plateaux with no visible higher peaks to aim for, rather than scaling
for ever upwards at a steady rate. Newton knew the speed of light, but
not all the implications of this - and it took centuries to get above
the plateau he conquered. I wanted to know if there are any indications
at all today that DNA science may go much higher yet, and it seems the
answer is No/We don't know/We don't think so/Ask again in 2120.

Peter Stewart
taf
2020-03-07 01:18:16 UTC
Reply
Permalink
Post by Peter Stewart
I wanted to know if there are any indications
at all today that DNA science may go much higher yet, and it seems the
answer is No/We don't know/We don't think so/Ask again in 2120.
Along the lines of your plateaus, sometimes they are eventually broken through, and sometimes they represent actually knowing what is happening. If we know something but don't know exactly why, our superficial understanding could be because of coincidence. As a trivial example, if you saw a coin right after it lands and record heads or tails, you may convince yourself that the process is random, when it is actually a fixed function of the physics applied to the coin: its weight, spin applied, upward momentum applied, etc., plus of its environment: air pressure and wind, temperature, distance to the landing site, etc., and if one identified and controlled all these factors, it would be clear our original understanding was completely off the mark. (I have often wondered if the randomness of nuclear decay might be something along those lines, only apparently random because we really are clueless about the underlying mechanisms - there is one documented case where a specific radioactive incident can be forced to be non-random, and though a special case it perhaps has broader implications regarding our understanding.) Gravity is another example, where we thought we knew what the rules were, had equations for it and everything, but not why. Then special relativity comes along and we realize what we thought we knew was just a superficial approximation of what was actually happening.

It is a little different when we know the mechanisms, like in genetics. We know how the DNA is replicated, how it undergoes meiosis and mitosis, how the chromosomal exchange works, how the inheritance of specific alleles lead to biochemical differences in the cell, etc., etc. While there are always refinements to this understanding, it seems to me a whole lot less likely with this level of detailed mechanistic comprehension that it would all just be a superficial similarity to what was really going on, of the type that would result in a plateau that would later be broken through. Short of descending into woo-woo land and positing that the biochemistry is actually all driven by psychic waves rather than physical forces, I don't see this understanding changing in terms of its big picture. I would expect a much deeper, more precise understanding in 2120, but not a completely different 'big picture'. The bars to using it for genealogical projects like your hypothetical are big picture kinds of issues, not precise detail issues.

Yesterday I thought through it a little and identified a handful of theoretical ways there could be something going on we don't fully understand that could potentially modify our model of DNA inheritance in a manner that would break through the roadblock that prevents identifying which grandparent DNA came from without further testing. Short of invoking invisible microscopic sentient beings inside the cells doing experiments on us like the mice is a Douglas Adams story, I couldn't make the hypothetical biochemistry of any of them actually work to give us the desired result. Of course, reality is not constrained by my imagination.

So, it would not surprise me at all if the big picture of quantum physics was completely different in 2120, but if we are talking about DNA inheritance, the only surprise I would anticipate would be if I was actually still around in 2120 to observe that there was no major change in our understanding.

taf
Peter Stewart
2020-03-07 01:34:01 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
I wanted to know if there are any indications
at all today that DNA science may go much higher yet, and it seems the
answer is No/We don't know/We don't think so/Ask again in 2120.
Along the lines of your plateaus, sometimes they are eventually broken through, and sometimes they represent actually knowing what is happening. If we know something but don't know exactly why, our superficial understanding could be because of coincidence. As a trivial example, if you saw a coin right after it lands and record heads or tails, you may convince yourself that the process is random, when it is actually a fixed function of the physics applied to the coin: its weight, spin applied, upward momentum applied, etc., plus of its environment: air pressure and wind, temperature, distance to the landing site, etc., and if one identified and controlled all these factors, it would be clear our original understanding was completely off the mark. (I have often wondered if the randomness of nuclear decay might be something along those lines, only apparently random because we really are clueless about the underlying mechanisms - there is one documented case where a specific radioactive incident can be forced to be non-random, and though a special case it perhaps has broader implications regarding our understanding.) Gravity is another example, where we thought we knew what the rules were, had equations for it and everything, but not why. Then special relativity comes along and we realize what we thought we knew was just a superficial approximation of what was actually happening.
It is a little different when we know the mechanisms, like in genetics. We know how the DNA is replicated, how it undergoes meiosis and mitosis, how the chromosomal exchange works, how the inheritance of specific alleles lead to biochemical differences in the cell, etc., etc. While there are always refinements to this understanding, it seems to me a whole lot less likely with this level of detailed mechanistic comprehension that it would all just be a superficial similarity to what was really going on, of the type that would result in a plateau that would later be broken through. Short of descending into woo-woo land and positing that the biochemistry is actually all driven by psychic waves rather than physical forces, I don't see this understanding changing in terms of its big picture. I would expect a much deeper, more precise understanding in 2120, but not a completely different 'big picture'. The bars to using it for genealogical projects like your hypothetical are big picture kinds of issues, not precise detail issues.
Yesterday I thought through it a little and identified a handful of theoretical ways there could be something going on we don't fully understand that could potentially modify our model of DNA inheritance in a manner that would break through the roadblock that prevents identifying which grandparent DNA came from without further testing. Short of invoking invisible microscopic sentient beings inside the cells doing experiments on us like the mice is a Douglas Adams story, I couldn't make the hypothetical biochemistry of any of them actually work to give us the desired result. Of course, reality is not constrained by my imagination.
So, it would not surprise me at all if the big picture of quantum physics was completely different in 2120, but if we are talking about DNA inheritance, the only surprise I would anticipate would be if I was actually still around in 2120 to observe that there was no major change in our understanding.
As someone who is mildly surprised to be still around in 2020, I wish
you every surprise in 100 years from now.

I what is known today covers all the basics of DNA inheritance and the
only blanks to be filled in are just colouring within an outline that is
already complete, then you have my permission to exhume my remains and
stuff a hat in my gaping mouth for my ghost to eat. I will be sure tell
John Schmeeckle how it went
Denis Beauregard
2020-03-07 05:55:56 UTC
Reply
Permalink
On Sat, 7 Mar 2020 09:36:49 +1100, Peter Stewart
Post by Peter Stewart
I am asking if there might ever be a possibility of determining that
position regardless of gender and without analysing DNA from all or even
most intervening generations.
As a project admin at FTDNA, I have observed the segments of
chromosomes shared by relatives. While you can't say what
part is from the father or the mother (we have pairs of chromosomes)
at this time (this will be possible later), it is easy to compare
grandchildren-grandparents or siblings. From observation, I can see
a chromosome is often splitted a small number of times. For example,
you would have (from grand-f-ather/m-other):

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmfffffffffffffffffffffffffffffffff

or

mmmmmmmmmmmmmmmmmmmmmfffffffffffffffffffffffffffffffffmmmmmmmmmmm

and not

mmmffmffmfffmmmmfmffmfmfmmfmmfmmfmffmfmmfmmmffmffmfffmmmmfmffmfmf

so it may be possible later to better identify the pattern of older
generations (but you will need many samples from the same family).

This will be possible when
- scanned segments will be long enough to separate father/mother
chromosomes in the results (I supposed this may be already done
with a whole genome, perhaps 150x, but you can't do 150x with
old DNA)
- scanning will be possible with almost no loss which is important
with old bodies where DNA may be hard to analyze

At this time, whole genomes are sold with coverages like 15x or 150x,
which mean they try to scan each segment of DNA 15 or 150 times. With
damaged DNA, there is just not enough DNA to do that.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
Peter Stewart
2020-03-07 06:34:09 UTC
Reply
Permalink
Post by Denis Beauregard
On Sat, 7 Mar 2020 09:36:49 +1100, Peter Stewart
Post by Peter Stewart
I am asking if there might ever be a possibility of determining that
position regardless of gender and without analysing DNA from all or even
most intervening generations.
As a project admin at FTDNA, I have observed the segments of
chromosomes shared by relatives. While you can't say what
part is from the father or the mother (we have pairs of chromosomes)
at this time (this will be possible later), it is easy to compare
grandchildren-grandparents or siblings. From observation, I can see
a chromosome is often splitted a small number of times. For example,
mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmfffffffffffffffffffffffffffffffff
or
mmmmmmmmmmmmmmmmmmmmmfffffffffffffffffffffffffffffffffmmmmmmmmmmm
and not
mmmffmffmfffmmmmfmffmfmfmmfmmfmmfmffmfmmfmmmffmffmfffmmmmfmffmfmf
so it may be possible later to better identify the pattern of older
generations (but you will need many samples from the same family).
This will be possible when
- scanned segments will be long enough to separate father/mother
chromosomes in the results (I supposed this may be already done
with a whole genome, perhaps 150x, but you can't do 150x with
old DNA)
- scanning will be possible with almost no loss which is important
with old bodies where DNA may be hard to analyze
At this time, whole genomes are sold with coverages like 15x or 150x,
which mean they try to scan each segment of DNA 15 or 150 times. With
damaged DNA, there is just not enough DNA to do that.
If quantum computing ever fulfills its promise, do you foresee that it
may be possible to make up reliably for at least some of the damage in
old DNA or are the permutations just too complex for this to be imaginable?

Peter Stewart
Denis Beauregard
2020-03-07 15:11:53 UTC
Reply
Permalink
On Sat, 7 Mar 2020 17:34:09 +1100, Peter Stewart
Post by Peter Stewart
If quantum computing ever fulfills its promise, do you foresee that it
may be possible to make up reliably for at least some of the damage in
old DNA or are the permutations just too complex for this to be imaginable?
From complete DNA, you can probably estimate 3 or 4 generations
back if you have enough people for analysis.

With old DNA, lets suppose you can do the same except you don't
know who is in the burial. So it is possible you may find with
more accuracy the relationship between 2 not identified bodies.

I think this is how improvements may help in medieval DNA.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
s***@mindspring.com
2020-03-07 19:15:12 UTC
Reply
Permalink
Post by Denis Beauregard
Post by Peter Stewart
If quantum computing ever fulfills its promise, do you foresee that it
may be possible to make up reliably for at least some of the damage in
old DNA or are the permutations just too complex for this to be imaginable?
From complete DNA, you can probably estimate 3 or 4 generations
back if you have enough people for analysis.
With old DNA, lets suppose you can do the same except you don't
know who is in the burial. So it is possible you may find with
more accuracy the relationship between 2 not identified bodies.
I think this is how improvements may help in medieval DNA.
I don't think that this answers the question that Peter asked. If quantum computing ever becomes practical, it will allow you to compute a huge number of possibilities at the same time, but it will not let you use data that isn't there. So, unless the degraded DNA has left traces that allow you to reconstruct its former sequence in a manner that is not currently understood (extremely unlikely, but see the comments at the bottom), the benefit of quantum computing will be the huge increase in the speed of analyzing hordes of data.

Reconstructing relationships from DNA is extremely difficult for all but the closest relationships. For example, using paper-trail matches with many of my DNA matches who happened to leave genealogical information, I have been able to determine which grandparent contributed which segments for more than half of my autosomal DNA with very high confidence, and for more than 90% of my autosomal DNA with reasonable confidence. For a large portion of these segments, I can determine which grandparent is the common ancestor for any matches having a large overlap on those segments, with the usual caveat that a small percentage of large matches are going to be false positives. There may be a small segment here and there where the evidence has misled me or I have misinterpreted it, but there are so many reinforcing pieces of evidence that I am confident that the picture as a whole is mostly correct.

However, if I do the thought experiment about how far I would have gotten if I had only the segment data without any of the paper-trail information, an entirely different picture emerges. I am lucky enough to have two siblings who have also tested, so using the Gedmatch website to analyze the places where we have no match, a single match, or a double match allows me to confirm that we are in fact full biological siblings, and to determine where most of the crossovers are, and in most cases who they belong to. The possibility of two crossovers being very close to each other can cloud the issue, but matches from more distant relations (when I could find them) would generally resolve that issue, even without paper-trail data. How much further could you get? In my case, not far, because I have no segment data for any relations closer than third cousin. If I had two matches on the same chromosome, I could often tell whether they were related through the same parent or different parents, but would generally have no information which was which, unless there was a sufficiently long match on the X chromosome, in which case the combination of matches between my sister, my brother, and myself would often tell which parent contributed the matching segment.

Suppose that I was also lucky enough to have a good number of first cousin matches (still without using any paper-trail data). In that case, there would most probably be enough matches on all of the chromosomes to separate which DNA belonged to which parent, with the X chromosome matches telling which parents were the mother and father. Having done this, I would have reconstructed a significant portion of my parents' DNA, so it is tempting to say that we just repeat this process ad infinitum.

But there are problems. First, I could only do this without paper-trail data if I also had some first cousin matches in addition to my sibling matches. Second, even if I could do this without paper-trail data, having data from three children of my parents would only allow me to reconstruct, on average, 87.5% of their DNA. (On average, it would be 50% for one child, 75% for two children, etc., with the missing part cut in half for each additional generation. The reality for a specific case could be far from the average.) Third, to carry the process further, I would have to do the same analysis for each of my close matches, hoping to reconstruct part of their parents' DNA before repeating the process. Thus, you have exponentially growing work for rapidly diminishing returns. After a small number of generations, the reconstructed DNA has thinned out so much that it is impossible to proceed further with any reliability.

Comparing the modern scenario to the scenario of using remains, the using remains scenario has two obvious disadvantages and one possibly useful advantage. The disadvantages are that the number of remains would almost certainly be smaller than the number of current DNA testees, and that the chances of getting usable DNA (and the amount of DNA obtained) would be lower. The advantage (unlikely to be relevant in most cases) is that after reconstructing somebody's DNA, they might be a close match to other remains (since we would not have the living people limit that modern tests have), which might just allow you to go further. So, if you happened to be lucky enough to have a set of remains that included related individuals from several consecutive generations, you might be able to reconstruct a pretty good tree. However, I would certainly not consider that to be a typical situation.

I will end with some pie-in-the-sky wishful thinking. Physicists believe that all information must be conserved at the quantum mechanical level. If this is not true, it would cause big problems for most current physics theories. On the other hand, scientists have no clue as to how such information might be reconstructed in any useful way. So, in theory at least, it would be possible to recover information from the past, even if we are currently clueless about how to do so. However, if this could be done, I would rather see a machine that could view past events than a machine that would reconstruct the past structure of partially recovered DNA. Now, that would be a really useful tool for genealogy (despite other concerns that such a possibility might raise).

Stewart Baldwin
taf
2020-03-07 19:32:15 UTC
Reply
Permalink
Post by s***@mindspring.com
I will end with some pie-in-the-sky wishful thinking. Physicists believe that all information must be conserved at the quantum mechanical level. If this is not true, it would cause big problems for most current physics theories. On the other hand, scientists have no clue as to how such information might be reconstructed in any useful way. So, in theory at least, it would be possible to recover information from the past, even if we are currently clueless about how to do so. However, if this could be done, I would rather see a machine that could view past events than a machine that would reconstruct the past structure of partially recovered DNA. Now, that would be a really useful tool for genealogy (despite other concerns that such a possibility might raise).
This reminds me of a discussion I once had with a physicist who claimed that since speech is nothing but waves of energy, and all energy is conserved in some form, it should be possible in some future where we are much more technologically advanced to measure all of the energy flux on earth at that time and reconstruct from it all of the wind and vibrations and waves, and extrapolate them back to the 18th century to learn the exact words that George Washington spoke when crossing the Delaware. I was skeptical.

taf
Peter Stewart
2020-03-07 21:25:59 UTC
Reply
Permalink
Post by taf
Post by s***@mindspring.com
I will end with some pie-in-the-sky wishful thinking. Physicists believe that all information must be conserved at the quantum mechanical level. If this is not true, it would cause big problems for most current physics theories. On the other hand, scientists have no clue as to how such information might be reconstructed in any useful way. So, in theory at least, it would be possible to recover information from the past, even if we are currently clueless about how to do so. However, if this could be done, I would rather see a machine that could view past events than a machine that would reconstruct the past structure of partially recovered DNA. Now, that would be a really useful tool for genealogy (despite other concerns that such a possibility might raise).
This reminds me of a discussion I once had with a physicist who claimed that since speech is nothing but waves of energy, and all energy is conserved in some form, it should be possible in some future where we are much more technologically advanced to measure all of the energy flux on earth at that time and reconstruct from it all of the wind and vibrations and waves, and extrapolate them back to the 18th century to learn the exact words that George Washington spoke when crossing the Delaware. I was skeptical.
Ah, but if a butterfly happened to flap its wings on the Congo at the
wrong moment, Washington's words on the Delaware might be misheard ...
Where is John Schmeeckle when we need him?

Thanks to both Todd and Stewart for taking the trouble and applying
their formidable skills in some of the most interesting and informative
posts I have read in 20+ years on this newsgroup.

I'm afraid over recent decades theoretical physicists have led each
other into a cul-de-sac of multiple "dimensions" and multiverses that
allow for more and more impenetrable models going nowhere at warp speed.
Common sense cannot take us very far in this direction (or after
milennnia we would already have the answers), but it should at least be
the departure point. We live in a universe with 6 (not 4, 10, 11 or 26)
observable dimensions of spacetime, if a dimension is understood to be
the encompassing condition within which scale can exist and not
vice-versa. Starting from there, who knows where we can end up?

Peter Stewart
Peter Stewart
2020-03-07 22:01:26 UTC
Reply
Permalink
Post by Denis Beauregard
On Sat, 7 Mar 2020 17:34:09 +1100, Peter Stewart
Post by Peter Stewart
If quantum computing ever fulfills its promise, do you foresee that it
may be possible to make up reliably for at least some of the damage in
old DNA or are the permutations just too complex for this to be imaginable?
From complete DNA, you can probably estimate 3 or 4 generations
back if you have enough people for analysis.
With old DNA, lets suppose you can do the same except you don't
know who is in the burial. So it is possible you may find with
more accuracy the relationship between 2 not identified bodies.
I think this is how improvements may help in medieval DNA.
I guess there must be some scope in future for triangulation with
identifying burials - for example, in sorting out which Bourbon may be
which or most closely related to which other from remains in Saint-Denis
that were scattered and mingled in the Revolution.

Peter Stewart
taf
2020-03-07 19:21:33 UTC
Reply
Permalink
Post by Peter Stewart
If quantum computing ever fulfills its promise, do you foresee that it
may be possible to make up reliably for at least some of the damage in
old DNA or are the permutations just too complex for this to be imaginable?
Computing power is not the limitation. Certain types of DNA damage just make it impossible to sequence.

If, for example, a base is knocked off the backbone (one of the most common types of damage), then all of the enzyme-based sequencing approaches available cannot sequence that piece of DNA beyond to point of the break. There are workarounds, but they invariably result in information loss or potential data corruption. There is a pore/electrical resistance based sequencing approach that is under development that might allow this type of damage to be jumped, but the data I have seen from it so far have been complete crap (with a known sequence, only one base in five matched the expected sequence), although this was several years back. People have also talked about electron-microscopy-based sequencing, but this has never come to anything, so it is a possibility, but to implement this type of sequencing for a whole genome would be prohibitively expensive and/or time consuming: high throughput long-read sequencing can give you a 'complete' genome in a couple of days, and short-read sequencing in about a week. In addition to the much more elaborate sample prep necessary, one would have to have a thousand cryo-EM machines at several million $ each (or different, more expensive devices) to get that kind of throughput, and the electricity needs would blackout the surrounding area. I can't see sequencing ancient DNA ever being that important (and I can't see this approach ever being developed to this scale because there are very few applications that would ever need this approach that could not be more simply done with others.

More commonly, though, the loss of the base causes a break in the whole chain, and that leaves an end that is particularly prone to chemical and enzymatic attack, causing even more lost information. If the damage is such that the DNA is chopped into pieces that are too short, there is not enough unique sequence remaining in the fragments to unambiguously relate it to a position on the genome (it would have an equal likelihood of going, say, any of 20 places that all have similar sequence, so it can't be placed at all). You couldn't even be able to tell if it was from the human buried or the bacterial that lived off his flesh after burial (although some bones are better than others at excluding such contamination, one of the bones of the inner ear is now preferred). These are just a few examples. Basically, DNA chopped up in too many pieces is a lost cause, no matter how good your computer.

The other thing that the damage does is it eliminates the possibility of detecting linkages. In a living person, you can get hundreds of millions of bases of contiguous sequence, particularly if you pair long-read sequencing for linkages with short-read sequencing for accuracy. That means all of the genetic markers on that contiguous read are known to come from the same parent. If you pair this with methylation analysis to detect parent-gender specific sites, you may be able to determine which parent the stretch came from. (This last is not actually done, but is theoretically possible.) Too many breaks in the DNA, and none of this can be done because the contiguous material is now on a thousand different pieces and you can no longer identify which pieces were once linked to each other and inherited together. Again, the breaks are resulting in loss of available data, much of which can't be compensated for with more computing power.

taf
Peter Stewart
2020-03-06 10:43:22 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered.
As a supplementary question to this - do you know of any studies in the
field of quantum biology that may point even vaguely towards "an
entirely unknown inheritance mechanism"?

Peter Stewart
taf
2020-03-06 12:45:25 UTC
Reply
Permalink
Post by Peter Stewart
Post by taf
Post by Peter Stewart
1. We have reasonable certainty in identifying the remains of
Charlemagne at Aachen and of the Merovingian king Childeric I at
Tournai. Supposing that adequate DNA can be extracted from both (which
as far as I know has not been tried), would you expect that one day
scientists may be able to tell with fair certainty if Charlemagne's
maternal grandfather was or was not a Merovingian?
As you specifically phrased the question, based on DNA only from these two burials, absolutely not, not now, not ever, unless an entirely unknown inheritance mechanism is discovered.
As a supplementary question to this - do you know of any studies in the
field of quantum biology that may point even vaguely towards "an
entirely unknown inheritance mechanism"?
Not that I am aware of.

I really just threw out that 'entirely unknown mechanism' because every once in a while a field is completely upended by a paradigm shift that so completely overturns our thinking that we find out everything we thought we knew was wrong (e.g. continental drift). Given the degree to which our conception of genetics conforms not only to observed inheritance, but also biochemistry, cell and molecular biology, the chances of this ever happening with genetics is minuscule, but scientists would have said something similar right before any of these paradigm shifts so I am not going to make the mistake of assuming we are not within one of these bubbles of misconception that could be burst by discovering something entirely beyond our current conception.

taf
s***@mindspring.com
2020-03-04 18:47:08 UTC
Reply
Permalink
Post by taf
The main problem is that most burials from the period cannot be identified by name, and those that can are not allowed to be studied. There are a few cases of a medieval burial being identified by DNA analysis, like in the Richard III case or, depending on when you draw the cutoff, Copernicus.
It is hard enough to come up with a medieval person with a known burial around whom there is a standing genealogical question. Something like the maternity of William Longespee, if that wasn't learned a couple of decades ago through traditional documentation. Then you would need a to be allowed to do the exhumation, unlikely if the purpose was solely genealogical curiosity. And for most genealogical questions, you would need a second test subject as well. It doesn't surprise me there have been no instances.
The concerns I raised earlier are an additional obstacle. The individual surname projects at Family Tree DNA have numerous (tens of thousands) of individuals whose 12 or 25 or 37 or 67 or 111 marker results are there for anyone to observe. In theory, you could use this data get a list of all of the individuals who match the Y-DNA results for Richard III most closely. It is extremely unlikely that any of these individual matches would prove anything, but there is a possibility that certain clusters of close matches might point the way to interesting lines of research which had not previously been obvious. It is also quite possible that such a search would strike out completely, but the lack of any reasonable search facility for all of these haphazardly posted results means that, in practice, we are currently denied the opportunity to try such a search to see whether or not anything interesting would emerge.

Stewart Baldwin
Denis Beauregard
2020-03-04 19:40:08 UTC
Reply
Permalink
Post by s***@mindspring.com
The concerns I raised earlier are an additional obstacle. The individual surname projects at Family Tree DNA have numerous (tens of thousands) of individuals whose 12 or 25 or 37 or 67 or 111 marker results are there for anyone to observe. In theory, you could use this data get a list of all of the individuals who match the
Y-DNA results for Richard III most closely. It is extremely unlikely that any of these individual matches would prove anything, but there is a possibility that certain clusters of close matches might point the way to interesting lines of research which had not previously been obvious. It is also quite possible that such a search
would strike out completely, but the lack of any reasonable search facility for all of these haphazardly posted results means that, in practice, we are currently denied the opportunity to try such a search to see whether or not anything interesting would emerge.

If you have a more or less complete signature for Richard III
(Y-37 ?), then it may be possible to at least identify a
descendant for a close match.

You may enter a part of the signature in google, like

"13 23 14 11 11-13 12 12 12 13"

" are important as this will search that exact series of numbers.

Then you may find either a match or a project with potential matches.

An admin can look for matches with a limited number of differences.
Having a nearly complete reference could lead you to a descendant.

I found an article about descendants of Louis XIII with the Y-37.
Perhaps there is something about Richard III or any other king
you may want to compare ?


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
taf
2020-03-04 19:49:21 UTC
Reply
Permalink
Post by Denis Beauregard
If you have a more or less complete signature for Richard III
(Y-37 ?), then it may be possible to at least identify a
descendant for a close match.
That is part of the problem. They only used a 'basic' test kit, so we really don't know enough for genealogy-level precision, just the millennia-old major branch. There was a proposal to do full genome sequencing, but I never saw whether it was funded, let alone producing results that might be more genealogically informative.

taf
Paulo Ricardo Canedo
2020-03-08 00:58:01 UTC
Reply
Permalink
Post by Peter Stewart
Post by Paulo Ricardo Canedo
Post by Peter Stewart
Post by Andrew Lancaster
Post by s***@mindspring.com
Post by Denis Beauregard
FTDNA has over 10,000 DNA "projects".
. . .
That is part of the problem. Suppose that there is a person of interest in one of those projects who has no (or very few) close matches within that project. There might still be a handful of exact (or very close) matches residing in other projects concerning unexpected surnames (or unexpected variant spellings) which might offer important clues on how to narrow the search. How do you find these matches (if they exist)? Some of the very large projects (such as the R-U106 project to which I belong) help with this to some extent, but most people who join one of the family projects do not bother to join the relevant large haplogroup project. Even in these large projects, the lack of search utilities means that there will be close matches lurking among the large miscellaneous groups that have not yet been placed in a narrower haplogroup, and these can be notoriously difficult to find without a search feature that allows you to find the closest matches of a specific testee.
The above concerns are not a problem for my own paternal line (at least as far as my own research goes), because I can get a list of my own close matches from the FTDNA site. However, if a direct male-line descendant of one of my female-line ancestors has had a Y-DNA test, there is no way (so far as I know) for me to search for his close matches. Some of the family project sites are useful, but one is often at the mercy of how competent the individual project managers are.
Stewart Baldwin
A lot of keen genetic genealogists help manage tests for relatives so that they can track more than just one male line.
Also see the reply of Dennis.
By the way it is also of course the case that the new autosomal tests are helping a lot of genealogists, including women. But except in cases where someone has tested ancient remains, this is not yet helping medieval genealogy as far as I am aware.
Can you given an example where someone has tested ancient remains and
you are aware that this has helped medieval genealogy?
Peter Stewart
Well, this is also not a specific genealogy but it's based on ancient remains. Some East Germanic remains have shown relevant amounts of East Asian DNA. This fits with East Germanics' ties to the Huns. What do you think of it?
If you are asking me, I think it may be interesting for historical
demographers but not specific enough to be helpful for medieval
genealogists. What I was asking for is a problem that occurs in the
study of medieval genealogy, not glancingly incidental to this, and that
has been resolved or at least clarified by testing ancient remains as
suggested upthread. I can't imagine an example, and apparently no-one
else here can either.
Peter Stewart
I think that, partially using that DNA evidence, we could plausibly argue for a descent of East Germanic Kings from Chinese Emperors, with the Huns and the Xiongnu as intermediate links, though, of course, it would be impossible to prove, let alone get specifis.
Denis Beauregard
2020-02-20 01:37:25 UTC
Reply
Permalink
Post by s***@mindspring.com
Its too bad that the tools for doing Y-DNA genealogical research are still so primitive (unless something has recently arrived on the scene that I didn't notice). The raw data available online comes from so many different sources (mostly family projects run largely by amateurs) that it is virtually impossible to run a decent
search for matches (except for your own kit on FTDNA) that any useful conclusions that might be obtained from the data are greatly diluted. I would love to see a reasonable GEDmatch-style utility for Y-DNA. I also wouldn't mind seeing something like that for mt-DNA, even though mt-DNA has limited usefulness as a genealogical
tool. (My own closest mt-DNA matches are "genetic distance" 2, hundreds, perhaps even thousands, of years beyond any documentable time-frame).

The fun of mtDNA matches ! You never know where is the MRCA.

I explored one case this afternoon. I wanted to find a 2nd match
for Julienne Dumont (her father is a d'Anglure, which will lead
to a royal descent if ever solved). So I checked the few matches
available. None shown her as the earlier known ancestor and most
had no online tree or only some hint to identify the parents (i.e.
the e.k.a. are the parents or name is not frequent). Nonetheless,
I was able to complete the lineage.

In another case, I have a complete triangulation with GD=3 or 4
and documented lineages for ancestress in the 1600s or 1700s.

While mtDNA (with 16,569 positions) can't be as useful as yDNA
(with over 1,000,000 positions), it may help to complete one lost
generation in a pedigree.


Denis
--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - http://www.francogene.com/gfan/gfan/998/
French in North America before 1722 - http://www.francogene.com/gfna/gfna/998/
Sur cédérom/DVD/USB à 1790 - On CD-ROM/DVD/USB to 1790
Peter Stewart
2020-02-16 21:29:30 UTC
Reply
Permalink
Post by taf
Post by Peter Stewart
I wonder if a DNA profile that might be meaningfully called "Islamic"
could be found by careful sampling in a place in the Arabian peninsula
away from the coasts and holy sites with repeated waves of immigrants,
visitors and pilgrims - for instance, by identifying people in the Ha'il
area with settled ancestry over the past 120+ years, from before Saudi
domination started changing the older social structure and demography.
But how far back this might go would still be problematic, since Arab
travellers and Islamic warriors who left the peninsula at different
times must have returned occasionally with half-alien offspring.
I think (going from memory), the genetic profile of the Arabian population from which Islam arose is pretty well understood - it is not dissimilar from that of Palestine at the same period, but with less 'Mediterranean' (e.g. Greek, Roman) influences. However, it would be problematic to use this as an 'Islamic' signature in the diaspora, where conversion and slavery played significant driving forces in the expansion. Of the non-native Muslim Iberians, Arabs formed an small segregated over-class, but demographically much more numerous were North African Berbers and Balkan slave-soldiers, and more significant yet were the muwallad native converts. Even with DNA this would be very difficult to sort out, though you could get a before-after comparison with enough sampling of burials from the post-Roman, pre-Islam era to compare with the modern inhabitants.
Less 'Mediterranean' and more 'African', I suppose, in the 7th century.
And given that not very much is definitely known about the extent of
trade and cultural exchange with the Indian sub-continent before then,
an ethnic snapshot of the earliest Muslim population - even if this
could be derived - would not be very distinct from surrounding peoples.

My suggestion about the Ha'il region was just based on a half-baked
notion that non-nomadic demographics might have changed less in that
long-settled area over the past 1,400 years than in other parts of the
peninsula. But the influence of Bedouin tribes would have tended to
homogenise DNA across a vast swathe of territory from modern Lebanon to
the Empty Quarter.

Peter Stewart
Paulo Ricardo Canedo
2020-02-20 11:16:37 UTC
Reply
Permalink
Post by taf
While the west-east grade is new, the presence of significant 'North African' DNA in Iberia has been known for a decade. However, this comes with two major caveats.
First, it can't be dated, and there is significant uncertainty whether it represents primarily the Berbers that accompanied or followed the Muslim invasion, or instead includes a significant contribution from two earlier periods, the Phoenician/Carthaginian settlement and significant cross-strait traffic during the Roman era.
Second, is the 'North African' DNA signal really North African? I am not up on the latest results for Iberia, but at least one model was that a single ancient population spanned the straits, but that in Iberia this signal was diluted by repeated subsequent influx while their signal still represents a more significant portion of modern North Africans.
The take-home - it is very difficult to correlate DNA signals with specific populations in a location with a history of repeated invasion and influx. A decade of extensive sampling was needed to definitively identify the genetic signal of the Indo-Europeans. It takes a fine-grained temporal and geographic analysis of ancient DNA from dateable burials before and after each population wave. I have not been following it of late (and haven't had time to read the new paper yet), but the best study I have read only used three time points separated by several thousands of years each (one anotu 15,000 ago, one 5000, and one Roman-era) which is insufficient to sort out all of the different migrations.
taf
It's really North African DNA, it matches present day North Africans, with their specific ancestral components.
I suspect the single ancestral population that you are refering to is Neolithic Iberian farmers. There's evidence that they colonized North Africa. However, that cannot be the explanation as our North African DNA also includes Iberomaurasian DNA.
taf
2020-02-20 17:18:13 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
It's really North African DNA, it matches present day North Africans, with
their specific ancestral components.
It is a long-standing error in these studies to equate current distribution with historical range, due to the degree to which population replacement has taken place. That it matches the genetic strain currently found in North Africa does not mean it was limited to North Africa in prehistorical times. To put this the other way, did this strain invade Iberia, or did sequential waves of Eurasian strains invade and mostly kill off/push out this strain, like what happened repeatedly in the rest of Europe? Either scenario would leave a strain in North Africa present in small amounts in Iberia.

taf
Paulo Ricardo Canedo
2020-02-21 18:48:45 UTC
Reply
Permalink
Post by taf
Post by Paulo Ricardo Canedo
It's really North African DNA, it matches present day North Africans, with
their specific ancestral components.
It is a long-standing error in these studies to equate current distribution with historical range, due to the degree to which population replacement has taken place. That it matches the genetic strain currently found in North Africa does not mean it was limited to North Africa in prehistorical times. To put this the other way, did this strain invade Iberia, or did sequential waves of Eurasian strains invade and mostly kill off/push out this strain, like what happened repeatedly in the rest of Europe? Either scenario would leave a strain in North Africa present in small amounts in Iberia.
taf
Could you, please, clarify me what cross strait population you are refering to? Are you refering to the Neolithic Iberian farmers? If so, they did colonize North Africa but that cannot be the explanation. Iberians, like me, have among the highest percentages of Early European farmer DNA in Europe, behind only Sardinia and possibly parts of Sicily.
In addition, as I already mentioned, our North African DNA also includes Iberomaurasian DNA and even a tiny percentage of Subsaharan African DNA. The latter one, BTW, is slightly higher in the South than in the North. I'm tempted to believe that the North African DNA in Southwest Iberians is from forcefully converted Muslims while the North African DNA in Northwest Iberians is from North African Christian refugees, who mixed less with Subsaharan Africans.
taf
2020-02-21 21:17:35 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
I'm tempted to believe that the North African DNA in Southwest Iberians
is from forcefully converted Muslims while the North African DNA in
Northwest Iberians is from North African Christian refugees, who mixed
less with Subsaharan Africans.
All I am saying is that it is all too easy to draw conclusions one is tempted to make. The field is full of examples of doing such, only to have later, better data, and in particular data from burials, demonstrate that these easy conclusions have been completely wrong.

taf
Paulo Ricardo Canedo
2020-02-20 11:06:35 UTC
Reply
Permalink
Post by j***@gmail.com
Post by Paulo Ricardo Canedo
This newsgroup has debated Islamic, Berber and Arab gateways in the Iberian peninsula. Well, this is not a specific link but DNA has shown that Portuguese, like me, and Spaniards do have North African DNA. It decreases from west to east. Western Iberians, like me, have between 5 and 13 percent. Central Iberians have between 2.5 and 5 percent. Eastern Iberians have between 1 and 2.5 percent. The Basques, who are not only from the far east of the peninsula but also from a mountainous region, have only between 1 and 2.5 percent. Strangely, there's not a North-South difference, with Northern Portugal (where I live), Galicia, the Asturias and Leon being among the regions with highest North African DNA. See https://www.eupedia.com/europe/autosomal_maps_dodecad.shtml#African for a map. Some of this North African DNA may have come from non-Islamic sources, though. Dear followers of the newsgroup, what do you think of all this?
First, I would say that there is no such thing as "north african DNA", except to mean "relatives of people who currently reside in northern africa". In simpler terms all you are saying is that people in western iberia are, on average, more closely related to people across the straight of Gibraltar than people in eastern Iberia. This is not surprising.
In even more simple terms, you are claiming that people, on average, are more closely related to each other the closer geographically they reside to each other. This is also unsurprising.
Since "Islamic" is a culture and religion, not a "race" of people, it is not really logical to try and figure out if the "northern african dna" was "islamic" or something else.
--Joe Cook
Actually, the fact that Western Iberians, like me, have mor North African DNA than Eastern Iberians is surprising, because Eastern Iberia was historically more integrated in the Mediterranean trade networks.
taf
2020-02-20 17:04:06 UTC
Reply
Permalink
I never really addressed some of this.
Post by j***@gmail.com
First, I would say that there is no such thing as "north african DNA",
except to mean "relatives of people who currently reside in northern
africa".
This may seem to be the case, but it isn't. When scientists involved in this type of study speaks of "North African DNA" they are not referring to people who currently reside in North Africa. Instead, this is a term of art used to describe the reconstructed genetic signature of the Maghreb in 'ancient' times. It is a distinct genetic 'strain' that only represents a portion of the genomes of modern northern Africans, who also have significant Arabian and sub-Saharan African DNA, the introduction of which mostly post-dates the 7th century.
Post by j***@gmail.com
Since "Islamic" is a culture and religion, not a "race" of people, it is not
really logical to try and figure out if the "northern african dna" was
"islamic" or something else.
There is nothing illogical about the question being asked, it is just an issue with the shorthand used to express it. The fully elaborated question is whether the presence of a significan North African DNA with a west-to-east gradient was introduced into Iberia as a consequence of the Islamic invasion, as opposed to representing more ancient population dynamics. Fully logical, fully valid question to ask.

taf
Paulo Ricardo Canedo
2020-02-16 11:53:43 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
This newsgroup has debated Islamic, Berber and Arab gateways in the Iberian peninsula. Well, this is not a specific link but DNA has shown that Portuguese, like me, and Spaniards do have North African DNA. It decreases from west to east. Western Iberians, like me, have between 5 and 13 percent. Central Iberians have between 2.5 and 5 percent. Eastern Iberians have between 1 and 2.5 percent. The Basques, who are not only from the far east of the peninsula but also from a mountainous region, have only between 1 and 2.5 percent. Strangely, there's not a North-South difference, with Northern Portugal (where I live), Galicia, the Asturias and Leon being among the regions with highest North African DNA. See https://www.eupedia.com/europe/autosomal_maps_dodecad.shtml#African for a map. Some of this North African DNA may have come from non-Islamic sources, though. Dear followers of the newsgroup, what do you think of all this?
Oops, for the Basques, I, actually, meant "between 0.5 and 1 percent.".
Paulo Ricardo Canedo
2020-02-17 08:42:46 UTC
Reply
Permalink
Regarding Peninsular Arabs, especially Bedouins, they have much more Natufian DNA than Levantine Arabs and much less Anatolian farmer and Iranian farmer DNA. I managed to model them in Global25 as 70% Natufian, 10% Anatolian farmer and 20% Iranian farmer.
Peter Stewart
2020-02-17 10:44:17 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
Regarding Peninsular Arabs, especially Bedouins, they have much more Natufian DNA than Levantine Arabs and much less Anatolian farmer and Iranian farmer DNA. I managed to model them in Global25 as 70% Natufian, 10% Anatolian farmer and 20% Iranian farmer.
How can you classify Bedouins as "Peninsula Arabs"? Do you mean some
tribe/s that did not belong to the larger confederations and limited
their movements to the peninsula, somehow finding adequate grazing there
year-round for their camels, goats and horses?

Most of the tribes moved annually into modern Syrian and Iraq, and many
of their leaders kept residences in Beirut from the early 20th-century
onwards.

Peter Stewart
Paulo Ricardo Canedo
2020-02-17 11:26:34 UTC
Reply
Permalink
Post by Peter Stewart
Post by Paulo Ricardo Canedo
Regarding Peninsular Arabs, especially Bedouins, they have much more Natufian DNA than Levantine Arabs and much less Anatolian farmer and Iranian farmer DNA. I managed to model them in Global25 as 70% Natufian, 10% Anatolian farmer and 20% Iranian farmer.
How can you classify Bedouins as "Peninsula Arabs"? Do you mean some
tribe/s that did not belong to the larger confederations and limited
their movements to the peninsula, somehow finding adequate grazing there
year-round for their camels, goats and horses?
Most of the tribes moved annually into modern Syrian and Iraq, and many
of their leaders kept residences in Beirut from the early 20th-century
onwards.
Peter Stewart
I used samples from Saudi Arabia that are available in Global25. BTW, the results of Bedouins and urban Saudis are very similar.
Peter Stewart
2020-02-17 11:54:52 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
Post by Peter Stewart
Post by Paulo Ricardo Canedo
Regarding Peninsular Arabs, especially Bedouins, they have much more Natufian DNA than Levantine Arabs and much less Anatolian farmer and Iranian farmer DNA. I managed to model them in Global25 as 70% Natufian, 10% Anatolian farmer and 20% Iranian farmer.
How can you classify Bedouins as "Peninsula Arabs"? Do you mean some
tribe/s that did not belong to the larger confederations and limited
their movements to the peninsula, somehow finding adequate grazing there
year-round for their camels, goats and horses?
Most of the tribes moved annually into modern Syrian and Iraq, and many
of their leaders kept residences in Beirut from the early 20th-century
onwards.
Peter Stewart
I used samples from Saudi Arabia that are available in Global25. BTW, the results of Bedouins and urban Saudis are very similar.
They are now, because Saudi domination has made it so. The Bedouins were
not natural recruits to Wahhabism, and their social structure was
disturbed by its imposition. In the late 19th century the townspeople of
the peninsula and the nomadic Bedouins mostly regarded each other as
alien, culturally and demographically. Kinship may have been
unscientific, but it was not by any means inclusive on either side.

Peter Stewart
Andrew Lancaster
2020-02-17 19:01:21 UTC
Reply
Permalink
Post by Paulo Ricardo Canedo
Regarding Peninsular Arabs, especially Bedouins, they have much more Natufian DNA than Levantine Arabs and much less Anatolian farmer and Iranian farmer DNA. I managed to model them in Global25 as 70% Natufian, 10% Anatolian farmer and 20% Iranian farmer.
Which kind of DNA?
Paulo Ricardo Canedo
2020-02-20 11:28:04 UTC
Reply
Permalink
Post by Andrew Lancaster
Post by Paulo Ricardo Canedo
Regarding Peninsular Arabs, especially Bedouins, they have much more Natufian DNA than Levantine Arabs and much less Anatolian farmer and Iranian farmer DNA. I managed to model them in Global25 as 70% Natufian, 10% Anatolian farmer and 20% Iranian farmer.
Which kind of DNA?
Autosomal DNA, of course. You can model Bedouins, Saudis and Yemenis, yourself, in Global25, if you want to.
Loading...