Roger Milkman

Electrophoretic Variation in Escherichia coli from Natural Sources

At each of five loci in 829 Escherichia coli clones from 156 samples from diverse natural sources, electrophoretic analysis reveals a prominent mobility class (frequency over 0.70) and 2 to 11 distinct mobility classes at lower frequencies. The frequency distribution of the classes argues against the importance of neutral mutations in allozymic variation. Heterosis is not the universal cause of genic polymorphism.

Further evidence of thermostability variation within electrophoretic mobility classes of enzymes

Two Drosophila melanogaster strains, each heterozygous for “fast” and “slow” alleles at the Adh locus, and each having balanced second chromosomes, were found to differ in the apparent thermostability of the slow allozyme. The two strains were crossed, and F1heterozygotes were separated on the basis of the origin of the slow allele. After electrophoresis, the cellulose acetate strips were treated 1 1/2 min at 35 C. The putatively more sensitive allozyme showed a strikingly greater response to heat. These findings further support the conclusion that electrophoretically cryptic allelic differences exist which are expressed in thermostability differences. Further application of this approach has revealed one similar sensitive slow allozyme and three cases of a relatively resistant fast ADH allozyme in wild-caught flies.

ALLOZYME VARIATION IN E. COLI OF DIVERSE NATURAL ORIGINS

ABSTRACT . The results of an electrophoretic analysis of allozyme (allelic isozyme) variation at 5 loci in E. coli from diverse natural sources are incompatible with the “neutral” hypothesis, that electrophoretic mobility variation is due largely to the random genetic drift of many adaptively equivalent alleles. The effective number of mobility classes is small, and they are distributed discontinuously over a considerable range. Neither the relationship of charge change to amino acid substitution, nor a sharp reduction in the number of possible equivalent alleles, nor any of the likely spatiotemporal population structures of the species E. coli can provide a way out of this conclusion.

Heterozygote excess due to population mixing.

Wahlund’s well-known principle (1928; see Li, 1969) describes a reduction in the frequency of heterozygotes (compared to expectations based on average allele frequencies) when populations differing in their frequencies of two alleles are combined. Li (1969) has pointed out that when more than two alleles are involved, not all heterozygote classes need appear deficient; indeed, an apparent excess can result when the frequencies of two of the alleles covary positively. Heterozygote frequencies in natural populations have elicited a great deal of interest recently, particularly in relation to possible mechanisms maintaining genic polymorphism, so it may be useful to have a simple expression for the maximum heterozygote excess that can be generated by the mixing of populations. For example, one allele may vary in relative frequency, while two additional alleles share the remaining relative frequency in a constant ratio. I n such a case, a population composed of individuals from two different source populations will tend to have an excess of heterozygotes involving the two latter alleles. While Li did not consider the case of a constant within-population ratio of frequencies, such cases may not be infrequent in nature. One example is that of the blue mussel, Mytilus edulis, whose North American populations have three common electrophoretically distinct alleles a t the leucine aminopeptidase locus. And while the frequencies of the LAP96 and LAPg8 alleles vary a great deal along the Atlantic Coast, they appear to remain in roughly the same ratio everywhere from Virginia to Maine. In two places each frequency (and thus their combined frequency) changes by a factor of two over a distance of a few miles (Koehn et al. 1975; Milkman & Beaty, 1970). Any two alleles, a, and a2, may be considered as a synthetic allele, a* = ‘al or a2’. If the frequencies of a, and u2 vary in a constant ratio, the Wahlund effect will be seen equally in all genotypes ala,, a,a2, a2a2 and a*a*. That is, in the event of population mixing, the frequency of each of these genotypes will exceed expectations by the same proportion. Thus, the familiar excess of homozygotes (a*a*, etc.) is matched by an excess of ala, heterozygotes. Let p , be the frequency of a* in a population, and p 2 be the frequency of a* in a second population. Let k = p1/p2. When the populations mix, let the ratio be m individuals from the second population for every individual from the first. In that case, barring selection and other supervening factors, the proportion of a* homozygotes will be

A Simple Exposition of Jensen's Error.

Jensen (1969, 1973) has used within-population heritability data to support his contention that IQ differences between races have a considerable genetic basis. Criticism of this reasoning has been frequent, but perhaps never categorical. A method is now described for illustrating the error simply and quantitatively. An example shows that high heritability of a property within each of two populations is consistent with a vanishingly small heritability in the combined population.