GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-04 > 1272204906


From: Jim T <>
Subject: Re: [DNA] 23andMe X-SNPs compared to FF
Date: Sun, 25 Apr 2010 07:39:20 -0700 (PDT)
In-Reply-To: <mailman.12093.1271879139.12642.genealogy-dna@rootsweb.com>


Jacques, On which one of your SNPs did Family Finder and 23andMe disagree?

I've also written programs to analyze X chromosomes. I've mostly used 23andMe results so far because more results are available for 23andMe than for FF. That may change in the future.

Writing programs that handle results from both labs simultaneously is possible, but is more complicated, so I only look at one lab's results at a time. I'm starting to switch from using 23andMe results to using FF results because FF has more SNPs in the regions of the X chromosome that I'm interested in. Which lab is best for your purposes depends largely on from which lab you can get the most results for the people who you are interested in. Some 23andMe customers may be hesitant to give you their results because their results contain medically significant SNPs.

One of my programs compares pairs of people with 23andMe X chromosome results and shows the locations of their blocks of 50 or more matching SNPs. I've just rerun it for the first time since Jan 2009 with data from Ben Moscia's spreadsheet and posted the results here:
http://tinyurl.com/32u8m7o

Jim Turner

> Hello
> (second trial, my prior message having been sent in error
> to genealog-dnas-request@)
>
> To what extent FTDNA (Family Finder) and 23andMe test the same SNPs ?
> I made the comparison for the X chromosome and here are some results.
> FTDNA
> YES NO Total by 23andMe
> 23&me YES 2,283 11,593 13,876
> NO 13,506
> Total by FTDNA 15,789
> 2,283 SNPs are tested by both companies (intersection).
> 23andme tests 11,593 SNPs which are not tested by FTDNA,
> and FTDNA tests 13,506 SNPs not covered by 23andMe.
> If you were tested by both companies, then these two independent tests
> now
> cover 25,099 SNPs of your X chromosome (in my case at least) and these
> two
> data sets can be easily merged.
>
> As concerns the results obtained at these 2283 common loci, as can be
> seen
> below, they are greatly in accordance in my case.
> There is one discrepancy for one SNP which was declared C by 23andMe and > T
> by FTDNA.
> Of the 2283, 479+636+607+457=2179 were identically identified by FTDNA
> and
> 23andMe;
> 95 loci which were *no-calls* for 23andMe were identified by FTDNA.
>
> FTDNA
>
> A G C T - Total
> 23&me A 479 0 0 0 2 481
> G 0 636 0 0 2 638
> C 0 0 607 1 1 609
> T 0 0 0 457 3 460
> - 46 7 8 34 0 95
> Total
> 525 643 615 492 8 2283
> BTW,
>
> rs4827527 (position 66246877) is not covered by FTDNA. This is the SNP
> which
> served as leader of the
> 25 successive SNPs which composed the various patterns named *EuroAsian
> majority*, *non Asian Minority*,
> the *Afro American Red*, &c.
> In case this message did not correctly showed up on your screen,
> the above comparison has been posted at
> [1]http://cerbere..ca/ftdnato23&me/comp-chx.html
> I am very interested in writing programs for DNA pattern recognition and
> datamining.
> For the moment I am hesitant about which database to use, FamilyFinder,
> 23andMe, or a merger of the two
> (or even + that of DecodeMe) ?
> Best regards
> Jacques Beaugrand,
> French Heritage DNA project
>[2]www.frenchdna.org
>




This thread: