Eve and her darn mtDNA

Glenn Morton (GRMorton@gnn.com)
Wed, 05 Jun 1996 22:23:17

Since Terry asked for a little less "origin's" stuff I will TRY to take a
break for a while since I am the worst offender (but this is not a

Terry and I disagreed about the continued viability of the Eve hypothesis.
I thought a few quotes (sorry Paul) from the works of Templeton might
shed some light on the issue. Alan Templeton was the biologist who
destroyed the Eve hypothesis.

The original Eve hypothesis claimed to prove that Eve was from Africa, and
lived around 200,000 years ago. This claim was based on the mistaken
belief that the analysis of the mtDNA lineages had found the lineage with
the least branches in the "family tree" (most parsimonious). The deepest
branches were supposed to tell the researcher where the family tree's root
was located. The original work had the deepest root in Africa.

The problem is that in a mathematical problem of this nature the number of
possibilities staggers the imagination. There is no way to examine all
possible solutions so shortcuts are taken. Templeton writes:

"For example, for the 147 haplotypes in Stoneking, Bhatia, and Wilson
(1986), there are 1.68 x 10^294 possible trees." Alan r. Templeton 1993,
"The "Eve" Hypotheses..." American Anthropologist 95(1); p. 52

"Maddison reanalyzed these data and found 10,000 trees five steps shorter
than the 'maximum parsimony tree' given in Cann, Stoneking, and Wilson
(1987). This set of more parsimonious trees contains two different basal
clades: the first consists of the seven African haplotypes ... and the
second consists of a mixed Asian-African clade." Ibid.

"In summary, the evidence for geographical origin is ambiguous. The
African-origin hypothesis was first postulated on the basis of an invalid
phylogenetic analysis of the restriction-site data and inappropriate
diversity measures, and the analysis of the sequence data repeated the
earlier errors." Ibid. p. 57

One of the problems was that the answer depended upon the order the data
was put into the program!!! Ibid. p. 52

"Hence the DNA sequence data are compatible with the common female
mitochondrial ancestor having lived any time between 102,000 years ago and
554,000 years ago under the assumption of a perfect molecular clock" Ibid
p. 59

Using another analytical approach yields an age range of 439,000 to
844,000 years ago for "Eve".

Templeton then goes on to point out the problems with believing a perfect
molecular clock.

He writes:
"In summary, the degree of uncertainty concerning the time of the deepest
root of the human mtDNA phylogeny is extreme. The narrow ranges given in
Stoneking, Bhatia, and Wilson (1986) or vigilant et al. (1991) do not
incorporate the most important sources of error, including evolutionary
stochasticity, a broad range of possible calibration rates, a bias toward
underestimation, and incontrovertible deviations from mutation-drift
equilibrium for Human mtDNA. Although there is uncertainty, the above
results indicate that the coalescent time is unlikely to be less than
100,000 years but that it is likely to be much older than 200,000 years."
Ibid p. 60

Most importantly, Templeton writes:

"In all of these studies, there is no doubt that the dendrogram has
African populations diverging first, but this pattern could arise either
from a population phylogeny or from the long-term patterns of gene flow
among human populations distributed throughout the Old World since the
time of Homo erectus." Ibid p. 67.

This is why paleoanthropologists are retreating from Eve. Oh yeah, Homo
erectus generally lived prior to 500,000 years ago.

Foundation,Fall and Flood