Re: When did "Eve" live??

Philip Deitiker (pdeitik@bcm.tmc.edu)
Thu, 02 Jan 1997 19:55:01 GMT

keith@pop.onwe.co.za (Keith) wrote:

>kmacke00@uoguelph.ca (Kevin Mackenzie) wrote:

>>Hello everyone,

>>A few years ago I remember hearing about a study (out of California I
>>think) in which the researchers determined (with reasonable confidence)
>>that all modern Homo sapiens have one common maternal ancestor. This was
>>based on a genetic study. Does anyone know the authors or title of any
>>publications on this study? Did the researchers estimate when this "Eve"
>>lived and does anyone know of a similar genetic study performed for male
>>ancestors?

>>Thanks in advance
>>Kevin.

>The original study was performed by Cann, Stoneking and Wilson (1987)
>in a paper entitled Mitochondrial DNA and Human Evolution, published
>in Nature, Vol 325, pp 31 - 36. There are others thereafter which are
>too numerous to list. It should be easy enough to find, but if you'er
>REALLY nice to me, I may post you a copy (I accept bribes). It was
>not definitive but basedon fairly good principles. Just be aware of a
>few flaws.

>The study was based on mitochondrial DNA (mtDNA). mtDNA is a remnant
>of DNA left in the mitochondria from when these energy houses of the
>cell were taken up as endosymbionts some ridiculous number of years
>ago. Mitchondria are only passed on from mother to child, no paternal
>mitochondria are inherited, therefore all mtDNA is of maternal origin.
>The only way to do a similar study with an 'Adam' would be to use the
>Y chromosome. Unfortunately this suffers from the problem of
>recombination, which mtDNA does not, and it would therefore be
>dificult to infer lineages from such data.

>The authors estimated 'Eve' to have originated somewhere in Africa and
>that she lived around 140 000 to 290 000 years ago. Unfortunately,
>several of the assumptions made to reach these conclusions were
>suspect. This is not to say that the work was all bad. I believed
>they were/are on the right track, but should have been more careful
>and perhaps taken longer to validate their data.

I would like to add to this that 'eve' is not a person but a set of
events which stipulate arguments about population size and history.
To generate 'eve' the authors have built a family tree of humantiy
based upon mitochondrial sequence. To reconstructure the events
leading to the tree, one looks at major branch points, number of
mutations between major branch points, etc and thus giving a fair
estimation of mutation rates then one approximates the events which
occur. Thus the eve argument is that in a reasonably small population
(say 1000 to 10000 peoples) a single female passed a mutant
mitochondria to her germ cell lineages. This mutant was passed to her
female offspring and for the next few thousand generations, while the
population was small, this mitochondrail genotype displaced (by mere
random chance) all others in the human popualtion. Likewise, this
_exclusive_ genetic drift event occurs slower in larger populations
and faster in small populations, the same is true of the Y chromosome
(in which only _certain_ regions can undergo recombination). So that
the bottom line of the 'eve' and 'Y' works is that the population of
humans, which eventually progressed to extant population, was small
(1000 to 10000 inds) for a very long period of time about 200KYA.
Because of the way the genetic drift occurs, this tree can only
predict the age of the most recent 'eve' event, with presumbably an
unknowable number of 'eve' events occuring prior to the most recent
exclusive drift event. Thus, the approximate 200KY estimate is a
minimal estimate of bottleneck age and duration, longer estimates
based upon mt composition are meaningless since it is beyond the scope
of predictability (unless basic premises such as mutation rates,
average reproductive age, etc are changed). The human genome may be
more useful because it is a diploid media (possibly giving a longer
window for examination); however, there are more idosyncratic
phenomena associated with chromosomal replication as apposed to
mitochondiral replication. Nonetheless, it should be possible to find
recombination resistance regions (as is seen in other organisms) such
that one can see whether exclusive genetic drift events have occured
within the last 400KY or not.

For a reference on the critique about the 'eve' work, there was an
article titled 'the myth about 'eve'' by Francisco Ayala in Science
from about 1.5 YA. This gives a better idea about how genetic drift
and population size recipricate each other. The conclusions of Ayala
on population age, OTOH, must be heavily criticized because his major
arguments are based upon HLA-DRB1 loci, and because of selective
pressure and potential for gene conversion (demonstrated in mice at
the equivilant locus (I-A/E) and in humans at an adjacent locus (for
HLA-B, see Parham and Ohta, Science 1996 or Watkins, Critical Reviews
in Immunology 1995, both MHC-I and HLA are part of my current project)
this loci is potentially the fastest evolving coding loci in humans.
So instead of his 69 alleles i would predict that during the
bottleneck period there were basically 14-16 alleles and this both
trims the minimal population size and minimal bottle neck age down
significantly from the estimates of ayala. BTW, currently there are
about 170 (not 69) known HLA-DRB1 alleles and the number could reach
into the thousands. This would mean (using this loci and average human
mutation rates) that the age of 'eve' population would be 10MY old
with a population size in the 10s of millions. Since this is hardly
likely, I would predict a better explanation for HLA diversity :^).

Philip