Out of Africa

L.A. Moran lamoran at gpu.utcs.utoronto.ca
Mon Mar 16 07:42:24 EST 1992


As one of the ones who started the debate last fall I want to assure the
other participants that I am far from bored. Keep it up! This is what
these newsgroups are for.

I appreciate the arguments over methodology. However, when I first read the 
Vigilant et al. PNAS and Science papers I did not attempt to repeat their 
analysis. Instead I noticed that many of the Kung! lineages differed by only 
one or two nucleotides. In my opinion it doesn't matter how few steps the 
final tree has as long as we are asked to believe that individual Kung! 
differing by such a small amount are so distantly related. It just doesn't
make sense to postulate that several of the deepest roots are between members
of such a small modern group. Doesn't anyone ever look at their data anymore
to see if the result is reasonable? 

The Science paper is a direct response to criticism of the earlier studies.
One would have thought that the authors would have been much more careful
in light of the fact that they knew their paper would be scrutinized. What
went wrong? Perhaps some of the authors could explain this to us.

The recent Wilson, Stoneking and Cann (1991) paper puzzles me. In that
paper they still maintain that all of the recent data on mtDNA is consistant 
with an African origin of modern humans. They claim that they welcome 
Maddison's (1991) criticism of Cann et al. (1987) but they do not concede 
that their conclusion was wrong. Instead Wilson et al. point out that a 
parsimony approach is "guaranteed to give the wrong answer in many cases 
because it ignores sample size". According to them the hypergeometric test of 
Vigilant et al. (1991) takes sample size into account and when applied to the 
results of Cann et al. (1987) it confirms an African origin. (Actually the 
test, according to them, confirms that a primary branch leads exclusively to 
African mtDNAs and this is taken to mean an Africn origin.)

Can someone knowledgeable explain this hypergeometric test to me? As far as 
I can see the analysis of mtDNA has not provided any scientific evidence in 
favor of an African origin of modern humans. 

Wilson, A.C., Stoneking, M. and Cann, R.L. (1991) Syst. Zool. 40, 363-365
Maddison, D.R. (1991) Syst. Zool. 40, 355-363.

Laurence A. Moran (Larry)
Dept. of Biochemistry               lamoran at gpu.utcs.utoronto.ca
Medical Sciences Blgd.                    (416) 978-2704
University of Toronto               FAX   (416) 978-8548
Toronto M5S 1A8, CANADA
 



More information about the Mol-evol mailing list