Archiver > GENEALOGY-DNA > 2010-03 > 1268184797

From: William Hurst <>
Subject: [DNA] FW: [ISOGG] Variability of mitochondrial DNA
Date: Tue, 9 Mar 2010 20:33:17 -0500
References: <15E4DC46C46F4B3A885386284C8A7065@NEWGAMES>,<BAY133-W8CB1175CD249BB94F7132F1340@phx.gbl>,<663228A683944D24BEF3A82C92EE7087@NEWGAMES>,<BAY133-W1093D2EA7BDB40CB4BF594F1340@phx.gbl>
In-Reply-To: <BAY133-W1093D2EA7BDB40CB4BF594F1340@phx.gbl>

Another one.

Subject: RE: [ISOGG] Variability of mitochondrial DNA
Date: Tue, 9 Mar 2010 12:39:44 -0500

Hi again Debbie and all,

For those of you who haven't looked, the very influential Dick Eastman said this morning: "If this proves to be true, it could again end the use of mitochondrial DNA for genealogy." I don't think so. I think the "discovery" of high rates of heteroplasmy in the Nature article may have more of an effect on forensic or medical uses of mtDNA than for genealogy.

For background on the random nature of mtDNA mutations, look at Ann Turner's article in JoGG from 2006:

The new study's "technology allowed Vogelstein and his collegues to find single-letter DNA changes that occur in as few as 1 in 10,000 mitochondria." This is the realm of microheteroplasmy, which I mentioned in passing in my JoGG article back in 2007; so even that's not new. In contrast, FTDNA uses a threshold of 30% for each variant before listing a heteroplasmy. Also, FTDNA disregards heteroplasmies when showing matches.

Having looked at hundreds of full-sequence mtDNA results, I can say that heteroplasmies are not uncommon. I haven't tried to arrive at a percentage. Heterplasmies in the HVR are obvious from looking at large mtDNA haplogroup projects. Look for mutations ending mainly in Y or R or N. Heteroplasmies tend to be relatively recent and are rarely found in defining mutations. However, back mutations in defining mutations are not that uncommon, and those must have gone through a heteroplasmic stage in past generations.

There is a new related article by Nicholas Wade in The New York Times at It has an interesting finding from the main article: "...the Johns Hopkins team realized there was much more variation than expected in healthy people's mitochondrial DNA - seven our of eight people had small proportions of variant mitochondrial DNA in addition to their main type." In my opinion, if a low-enough threshold is used, as here, everybody would start to show heteroplasmies at the variable sites, defeating the purpose of the testing. Whether FTDNA's 30% is the optimal threshold or not, I don't know; but it appears to work for our purposes. Also, all FTDNA test results, at least, are based on the same type of tissue samples from cheek swaps, avoiding the major issue found in the Nature study.

Wade also quotes a Terry Melton to the effect that this new "discovery" is not that new. Melton had an interesting paper on this back in 2004:

Back to the blog from The Scientist which I posted before, there is an interesting comment by Hugh Fletcher which says in part "The transport of female slaves around Europe, Africa, Asia for millenia have ensured that any mitochondrial type could appear almost anywhere." I've seen that argument before, that mtDNA is too spread out geographically to be of any use in tracing ancestry. I've seen enough geographically specific subclades forming out of FGS results to know that this theory is not true.

So, this new paper - which I haven't seen in full - appears to advance the science; but it is not as earthshaking as some such as Dick Eastman may think.

Bill Hurst

This thread: