GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2009-09 > 1254148809


From: David Faux <>
Subject: Re: [DNA] 23andMe and Proving Native Ancestry
Date: Mon, 28 Sep 2009 07:40:09 -0700
References: <ea3bd9560909261607g40841da4h38d051beea026eef@mail.gmail.com><425095.82468.qm@web24507.mail.ird.yahoo.com>
In-Reply-To: <425095.82468.qm@web24507.mail.ird.yahoo.com>


Anders,

I will be very interested to see what type or software / algorithm can
determine anwers to the matters below. I have assumed that the only way to
ascertain whether my Xibo - Yakut block is from my NA ancestor is to go
about collecting samples from those who are band members of various First
Nations groups residing in the region of the Great Lakes; At the moment the
block in question is infered to be NA since it is precisely the size
expected based on paper genealogy - but the matches are to NE Asians (with
the strongest genetic link to Native Americans). The block also has a start
point not seen in any Europeans, only a few East Asians, but a long list of
Native Americans from HGDP-CEPH. All this strongly suggests a NA connection
but it only comes into focus at this point because it is consistent with
predictions from standard genealogy of an X with 6.5% NA..

There is now some discussion about the C3 group (Y-DNA) having arrived in
the Central Plains much later (c. 8 k ybp) than the Q dominated initial
group (c. 14 k ybp).
The entire migration situation may be much more complicated or different
than our conceptualizations to date. It appears by the archaeological and
morphological record that in North American the C3 group (e.g., Plano group)
may have largely replaced the earlier group (e.g., Clovis group). They
likely brought autosomal and x sequences that are not today found in South
America. Hence the only way we are going to see them, if there was a
founder effect involving a second migration that did not reach South
America, is by collection of reference samples from North America.

There is also the region flanking the Xibo block toward the telomeric area.
As you know it is about 12 Mb and does not match any group, no matter how
low you set the bar. It would make sense that this is a rare block, due to
a founder effect and like mtDNA X2a is perhaps only to be found in the Great
Lakes region - not elsewhere in the Americas or in Asia. If found in NA of
the central regions of North America only, then as with the otherwise
mysterious X2a, it is NA. I don't see how one could "unearth" this data any
other way than collecting reference samples from the few regions where it is
likely to be found.

Anyway, I have learned that you are developing an expertise in this area
that leaves almost all of us "in the dust" (a pioneer blazing new paths and
all that). So I guess we will see what your analyses show and go from
there.

David K. Faux.

On Mon, Sep 28, 2009 at 6:25 AM, Anders Pålsen <> wrote:

> David, list.
>
> It may not be absolutly neccessary to have more native american reference
> samples from the Mohawk area or geographical close region. I am currently
> investigating locus ancestry software to help solve this kind of problems
> where it should be possible to infer ancestry on segments of the chromosomes
> by clustering analysis of linked SNPs in a chromosome segment.
>
> I am hoping to be able to infer as a alternative and supplement to the
> shared segment analysis between individuals in PLINK the ancestry for both
> identified shared segments and empty spaces without any shared segments
> based on clustering because clustering do not need to have long segments of
> identical SNP (but similar enough to cluster) and this way able to identify
> the ancestry of the segment even there is no shared segment matches in a
> PLINK analysis.
>
> This is actually what happend in David's case, he first identified the
> "east-asian" segment on X-chr by looking at deCODEme's locus ancestry graph
> for himself on the chromsome, this segment showed up because it clustered to
> the chinese and japanese HapMap panel even it maybe was no long identical
> segment match to any of these. I later identified the specifics and shared
> segment in PLINK to a few specific north-asian individuals in the HGDP
> panel. Currently the deCODEme and 23andme locus ancestry analysis cant
> differenciate the origin into east-asian or native american cluster, but
> only by proxy to east-asians in HapMap, however as David mention above his
> Xibo shared segment there is no identified PLINK shared segment with anyone
> even when setting the bar to >25 SNP and >1 Mb in PLINK, but if we look at
> his deCODME locus ancestry graph from deCODEme the HapMap data suggests that
> this region is of european ancestry. It may of course be so but at the same
> time when looking at the deCODEme genome browser collection of native
> american individuals you will find a larger minority part of "european"
> ancestry even among the apparent most isolated unadmixed tribes from the
> HGDP panel, so the deCODEme analysis do open for classification errors in
> this case wrongly assuming native americans = east asians, some true native
> american segments may look "european" and cluster with "europeans" in a wide
> sense without been the result of recent admixture but from ancient
> central-asian "european" ancestry of native americans, not to mention the
> centromere related erronous "african" match David have mentioned earlier a
> direct result of limited population data in deCODEme locus analysis.
>
> The above mentioned no PLINK sharing segment matching but clear locus and
> total genome clustering you would probably observe if you have unrelated
> parents from the same area or etnicity and do a friends sharing analysis of
> them in 23andme or deCODEme. They share few smaller segments and the local
> ancestry graph can vary greatly between them but their total ancestry
> estimate look very similar. So in this kind of locus ancestry analysis it
> may be a advantange to have a panel ancestry informative SNP for the locus
> ancetry analysis. There are several studies of autosomal AIM SNP available
> but I have not managed to find any collection of X-chromosome AIM SNP
> research papers so it may be a task I must do on my owm. If anyone have seen
> these kind of studies email me.
>
> Anders
>
>
>


This thread: