Xuhua Xia says
>There is an interesting feature of any multi-level control system in which
>the successful completion of events at one level depends on successful
>completion of previous levels. In terms of the ontogenic chain, we can label
>the probability of successful completion of developmental processes
>controlled by Group 1 genes as P1, the probability of successful completion
>of developmental processes controlled by Group 2 genes as P1*P2, ..., and
>the probability of successful completion of developmental processes
>controlled by Group N genes as
>> P1*P2*...*P(n)*...*P(N).
>>Thus, the probability of successful completion of developmental processes
>decreases with n.
I am not a developmental biologist, but it seems to me that this idea is
quite old. Did C.H. Waddington not present arguments along similar lines?
The weakness in the probability model above is the fact that the "success"
of each process is assumed independent of the others, apart from out and
out failure. Is it not possible that an individual gets off to a shaky
start (developmentally speaking)and completes the processes associated with
P1, but its weakened condition increases the likelihood that P2 is not
completed (i.e. causes an increase in 1 - P2). It seems likely that many
genes in different "groups" may have such effects on one another and
alter the probability of the "successful completion" of the associated
developmental processes. The "independence" assumption for gene groups
then seems rather questionable.
Bruce Rannala
Department of Biology
Yale University
