Conclusions and Discussion - Genetic Diversity in Establishing Plant Populations: Founder Number and Geometry - page 328

“genetic drift is a more powerful force for altering gamete frequencies at the
multilocus level than at the single locus level.” What is the relevance of the
previous statement in light of the genetic processes we have demonstrated
with the previous NEWGARDEN analyses?
We have shown that, all other factors held constant, differing
arrangements of founders between two founding events can cause low-
frequency unique alleles to be lost at different rates from the resulting
two newly expanding isolated populations. As unique alleles are lost at a
locus, other alleles at the same locus increase in relative frequency. Now
consider a locus for these same two populations that has only a few alleles,
all at relatively high frequencies. Just as low-frequency alleles are being
lost at different rates in the former situation, when alleles are at higher
frequencies, copies of those alleles are being lost at differing rates in the
two populations as a consequence of the contrasting founding geometries,
especially when population numbers are low. Even though unique alleles
are not lost at these loci from the latter two populations, their alleles will
shift in frequency at different rates as copies of alleles are lost at different
rates. When copies of an allele are lost at higher rates, the total frequency
of the alternate alleles at that locus will increase (although this increase can
be distributed across the alternate alleles in different ways across replicate
trials). Thus, a given unique common allele may randomly increase or
decrease in frequency because of drift caused by spationumeric founding
effects. Of our two newly founded populations, the one losing the greatest
number of low-frequency unique alleles should also be losing the greatest
number of copies of more frequent alleles. This greater loss of copies at
higher-frequency alleles will lead to increased variance in their frequencies
in replicate comparative trials, this increased allele frequency variance
stemming from the same spationumeric effects that drive increased loss of
low-frequency unique alleles.
These drifting changes of unique allele frequencies (including the loss of
rare alleles) will alter the probabilities of occurrence of particular multilocus
combinations of alleles in the gametophyte and sporophyte generations. To
give just one simple example, imagine three loci (call them loci 1, 2, and 3)
that assort independently. These loci interact in a multilocus manner to affect
a trait. Further, let us consider just one distinct allele for each locus, calling
these alleles 1d, 2g, and 3c at these 3 numbered loci, respectively. Suppose we
have two populations (call them X and Y) both with the same low number
of founders, and the two sets of founders are genetically identical, including
the frequencies of alleles 1d, 2g, and 3c, as shown in Table 18.1. Imagine that
all initial conditions are held constant except that the founders establish
in different geometries and, as a consequence, as the two populations
develop they have different rates of losses of alleles similar to differences
demonstrated in NEWGARDEN analyses outlined previously in this topic.