Hill-Robertson Interference Reduces Genetic Diversity on a Young Plant Y-Chromosome

Abstract

X and Y chromosomes differ in effective population size ( ), rates of recombination, and exposure to natural selection, all of which can affect patterns of genetic diversity. On Y chromosomes with suppressed recombination, natural selection is expected to eliminate linked neutral variation, and lower the of Y compared to X chromosomes or autosomes. However, female-biased sex ratios and high variance in male reproductive success can also reduce Y-linked , making it difficult to infer the causes of low Y-diversity. Here, we investigate the factors affecting levels of polymorphism during sex chromosome evolution in the dioecious plant (Polygonaceae). Strikingly, we find that neutral diversity for genes on the Y chromosome is, on average, 2.1% of the value for their X-linked homologs, corresponding to a chromosome-wide reduction of 93% compared to the standard neutral expectation. We demonstrate that the magnitude of this diversity loss is inconsistent with reduced male caused by neutral processes. Instead, using forward simulations and estimates of the distribution of deleterious fitness effects, we show that Y chromosome diversity loss can be explained by purifying selection acting in aggregate over a large number of genetically linked sites. Simulations also suggest that our observed level of Y-diversity is consistent with the joint action of purifying and positive selection, but only for models in which there were fewer constrained sites than we empirically estimated. Given the relatively recent origin of sex chromosomes, our results imply that Y-chromosome degeneration in the early stages may be largely driven by selective interference rather than by neutral genetic drift of silenced Y-linked genes.

The effects on neutral variability of recurrent selective sweeps and background selection

“…http://dx.doi.org/10.1101/358309 doi: bioRxiv preprint first posted online Jun. 29, 2018; predicted relative diversity value for a 70 gene region is quite close that observed for the fourth chromosome of D. melanogaster, which has a comparable number of genes , suggesting that diversity in non-crossover regions of the genome is largely controlled by BGS, as was also inferred by Hough et al (2017) for the case of the newly evolved Y chromosome of Rumex.…”

“…This is what is seen in most of the results for autosomal loci shown in Table 4 (Table S4 of File S1 presents some parallel results for X-linked loci). The most extreme case is when there is no crossing over, a regime in which the efficacy of BGS is undermined by Hill-Robertson interference among the deleterious mutations, so that the assumptions underlying the BGS equations tested in the previous section do not apply (McVean and Charlesworth 2000;Comeron and Kreitman 2002;Kaiser and Charlesworth 2009;Seger et al 2010;Good et al 2014;Hough et al 2017). For the other rates of crossing over, there is much closer agreement between the two estimates of B, although the value for favorable mutations is nearly always larger than for neutral mutations.…”

Ancestral and neo‐sex chromosomes contribute to population divergence in a dioecious plant

“…This problem is likely to be especially acute on the Y chromosomes, as linked selection has been shown to have a strong effect on genetic polymorphism, including in R. hastatulus where diversity at Y‐linked loci has been found to be 2.1% that of their X‐linked homologs, corresponding to a 93% reduction compared to neutral expectations (Hough et al. ). To test whether high F ST values were caused by low levels of within‐population polymorphism, we calculated Nei's d XY .…”

“…(), were categorized as autosomal, following Hough et al. (). Haplotypes were identified as X or Y depending on sex‐specific SNPs, following Hough et al.…”

“…Haplotypes were identified as X or Y depending on sex‐specific SNPs, following Hough et al. () using find_perfect.pl: briefly, reference alleles heterozygous in all males, and homozygous in all females were identified as X‐linked, whereas the non‐reference allele, absent in all females, was identified as Y‐linked; this method is contingent on a female transcriptome assembly as the reference during alignment.…”