A model of electrophoretic variation: Finite populations and finite numbers of mutable sites (Q1064282)

In this paper we present a model of electrophoretic variation for a protein with M amino acid positions, each of which can be in the selectively equivalent charge states 0 and 1. At each amino acid position the mutation rate from one charge state to the other is \(\nu\). We derive the expected variance in charge, for a population of N individuals in generation t, which is \(\sigma^ 2(t)=[2MN\nu /(1+8N\nu)]\{1-\exp [- (4\nu +1/(2N))t]\}\) for an initially monomorphic population. This variance, which is unaffected by intragenic recombination and gene conversion, does not increase linearly with N as does the steady-state variance in charge for the older electrophoretic models. We also derive the probability, \(P_ j(t)\), that a random gene in a population exceeds another by j charge units. We conclude: (1) that, except for species which have maintained effective population numbers in the millions, the infinite mutable sites model adequately describes neutral genetic variation; (2) that very high levels of 'neutral' genetic variation can be attained only by genes with a fairly large number of 'neutral' amino acid sites in species which have maintained an effective size in the millions for millions of generations; (3) that, because all species surveyed have a mean heterozygosity of less than 25\%, it is likely that most genes have only a few 'neutral' covarions.