Genetic diversity and demographic history
Levels of genetic diversity were calculated in 10kb windows separately
for eastern and western monarchs. Genomic windows were classified into
autosomes, Z chromosome and neo-Z chromosome. Levels of genome-wide
genetic diversity of eastern and western monarchs were essentially
identical (Fig. 6A, S8, Tables 1, S6), with the genome-wide genetic
diversity landscape of eastern and western monarchs showing an almost
perfect correlation (Fig. S8). This provides further evidence for a lack
of genome-wide genetic differentiation. The Z chromosome had a lower
nucleotide diversity than the autosomes, and the neo-Z chromosome had an
intermediate nucleotide diversity, reflective of their effective
population sizes (Table 1). The
neo-Z chromosome was identified to be an ancestral autosome but fused to
the Z chromosome in the monarch butterfly
(Gu et al. , 2019). The dosage
compensation of the neo-Z chromosome segment looks more similar to the
autosomes than the ancestral Z chromosome
(Gu et al. , 2019). Consistent with
this finding, we found that the genetic diversity of the neo-Z
chromosome segment is higher than the ancestral Z-chromosome (Table 1).
In line with genetic diversity (θπ ), genome-wide
Tajima’s D was also highly similar in eastern and western monarchs
(Table 1; Fig. 4, 6B) suggesting that eastern and western monarchs have
a similar demographic history. The negative genome-wide Tajima’s D in
both eastern and western monarchs indicates a recent recovery from a
past bottleneck. We found a total of 128 common Tajima’s D outlier
windows between eastern (out of 223 windows) and western (out of 202
windows) monarchs (Table S7). We did not identify any common windows
between the Tajima’s D outliers and genetic differentiation outliers,
suggesting that the low genome-wide genetic differentiation is the
effect of consistent gene flow between these two groups rather than
random genetic drift.
We used ∂a∂i to scan the likelihoods of 15 models in total to find the
best demographic model to explain the 2D-SFS of eastern and western
monarchs (Fig. S9). A summary of the likelihood scores of the optimized
models is given in Table S8. An extended table with the best 5
iterations in each model and their optimized parameters is given in
Table S9. The model “bottlegrowth_split_mig” showed the highest
log-likelihood score and the lowest Akaike Information Criterion (AIC)
(Fig. 7, S9, Table S9). This model assumes a bottleneck before
divergence between the two populations, followed by an exponential
growth in population size with migration. Top 5 iterations of
“bottlegrowth_split_mig” gave consistently higher likelihood scores
than all other models. Both “no divergence” and “no migration”
models had low likelihood scores in the optimization. Models with
bottleneck and exponential growth before the split showed the highest
likelihood scores, giving the weight to this scenario (Tables S8, S9).
The model “bottlegrowth” which considers a bottleneck followed by
exponential growth without a split into two populations had low
likelihood (Fig. S9, Tables S8, S9). We visually verified the most
likely model by plotting data and model outputs using the “Plot_2D”
function in ∂a∂i (Fig. S10). We used the parameters of the iteration
with the highest log-likelihood score for the model
“bottlegrowth_split_mig” to calculate effective population sizes,
migration rate and time of the bottleneck (Appendix 1 in Supplementary
Materials). According to these parameter estimates, eastern and western
monarchs experienced a bottleneck about 412 thousand years, underwent
exponential growth, then diverged about 112 thousand years ago with a
migration rate of 7.33*10-07 per generation (Fig. 7).
∂a∂I analysis also showed that eastern and western monarchs have similar
effective population sizes with a symmetric migration (Fig. 7, S9, Table
S6, Appendix 1). This finding was also confirmed by the Tajima’s D
values calculated across the genome in windows of 10kb. Tajima’s D was
similar for eastern and western monarchs (Fig. 4, 6B, Table 1).