Published allele frequencies of coat genes from 204 cat (Felis catus)
populations all over the world are studied by means of correlation and
principal component analyses. Locally associated groups of alleles wi
th correlated frequencies are revealed [the (O, d, I, W), (a, t(b)), a
nd (S) allele groups]. The phenomenon must have resulted from the para
llel action of local selection vectors on a worldwide scale. In partic
ular, the (O, d, I, W)-group is formed by the vector of artificial sel
ection. It is shown that principal component analysis facilitates reve
aling of such associated alleles connected by moderate correlation. Th
is analysis also allows us to classify the populations by plotting the
m in two-dimensional space of two main components containing the most
part of the total allele frequency variance.