Perennials - Genetic Diversity in Establishing Plant Populations: Founder Number and Geometry - page 239

1. Greater loss of unique alleles near borders increases inbreeding: If this
factor contributes signifi cantly to higher F values when founders are
placed near borders, we would, fi rst, expect signifi cant differences in
either expected or observed heterozygosity values for trial v versus
trial W, which is not the case ( Fig. 15.9B) . Further, we would expect a
pronounced increase in F for trial v relative to trial W due to inbreeding.
In fact, the reverse is true (Fig. 15.10A) . Finally, there should be greater
loss of unique alleles in trial v than in trials W, and also in trials m
compared to trial M, which is not supported by the data (Fig. 15.10B).
These results suggest that the increased F values for trial m versus M
do not result from a greater loss of alleles in trial m due to the greater
proximity of founders to a border compared to trial M. Increased loss
of alleles near borders leading to more inbreeding does not appear to
be a major contributing factor to the increased F values for trial m.
2. The contribution of Wahlund effect: The difference between expected
and observed heterozygosity should increase when two subpopulations
are analyzed as though they are one. Comparisons for expected versus
observed heterozygosity across these four trials are shown in Fig.
15.9B. In both trials m and M, NEWGARDEN analyses are calculated
as though they are one population even though trial m is subdivided
into isolated subpopulations. Note that the subdivision does not affect
expected heterozygosity values in trials m versus M as shown in Fig.
15.9B. The effect of this pattern on Wahlund effect is discussed in the
next section.
3. Increased inbreeding rates in smaller groups: In the same graph
(Fig. 15.9B), note that observed heterozygosity (actual counts of
heterozygotes) is much lower in trial m (founders subdivided) than
in trial M. Further, the observed heterozygosity value for trial m is
similar to the values for trials v and w (both also with isolated groups
of 43 founders). Under item (1) above, we established that loss of
heterozygosity stemming from increased inbreeding resulting from the
loss of alleles at borders is not a signifi cant factor (compare observed
heterozygosity for trials v (founders near border) versus W (founders
central)). That losses of heterozygosity are very similar when small
groups of 43 founders are used, whether there is only one group (trials
v and V) or four such groups (trial m), suggests that placing founders in
small groups leads to greater rates of loss of observed heterozygosity.
The slight increase in observed heterozygosity of trial m relative to trials
v and W is likely due to increased gene fl ow among subpopulations
over generations. In all cases where small groups are involved (trials m,
v, and W), observed heterozygosity is reduced whether near borders or
not, relative to when founders are placed in one large group (trial M).
The greatest decrease in observed relative to expected heterozygosity