Figure 2.

Fixation probabilities estimated from W–F simulations plotted as a function of inversion length for autosomal (A and B) and SLR-expanding inversions on Y chromosomes (C and D). Point shapes indicate different chromosome-arm-wide mutation rates relative to selection (i.e., different values of $U$), which influences the average deleterious mutation load carried by a standard-arrangement chromosome ($U/\left(hs\right)$). Dashed horizontal lines indicate the corresponding expected fixation probability for a neutral variant for the same population size, and hence correspond to values of $\mathsf{\text{1/}}\left(\text{𝟤𝑁}\right)$ for autosomal inversions, and $\text{𝟤}/N$ for Y-linked inversions. In (C and D), we show fixation frequencies for all inversions (grey points), for only initially unloaded inversions (red points; i.e., the $d=\text{𝟢}$ inversion class), and the analytic approximation given by multiplying the probability that an inversion is initially mutation free by the effective initial frequency (unfilled points; i.e, $\text{𝖯𝗋}\left(\text{fix}\mid \text{𝑥}\right)\approx {\text{𝑓}}_{\text{𝟢}}{\text{𝑞}}^{\ast }$). Note: red points almost perfecty overlay grey points, but the few small discrepancies reflect rare additional fixations of initially (lightly) loaded inversions (where $\text{𝑑}>\text{𝟢}$). Other parameter values were set to: $h=\mathsf{\text{0.25}}$, $s=\mathsf{\text{0.01}}$, and $n_{tot}={\text{𝟣𝟢}}^{\text{𝟦}}$.