The effective population size modulates the strength of GC biased gene conversion in two passerines

Understanding the determinants of genomic base composition is fundamental to understanding genome evolution. GC biased gene conversion (gBGC) is a key driving force behind genomic GC content, through the preferential incorporation of GC alleles over AT alleles during recombination, driving them towards fixation. The majority of work on gBGC has focussed on its role in coding regions, largely to address how it confounds estimates of selection. Non-coding regions have received less attention, particularly in regard to the interaction of gBGC and the effective population size (Ne) within and between species. To address this, we investigate how the strength of gBGC (B = 4Neb, where b is the conversion bias) varies within the non-coding genome of two wild passerines. We use a dataset of published high coverage genomes (10 great tits and 10 zebra finches) to estimate B, nucleotide diversity, changes in Ne, and crossover rates from linkage maps, in 1Mb homologous windows in each species. We demonstrate remarkable conservation of both B and crossover rate between species. We show that the mean strength of gBGC in the zebra finch is more than double that in the great tit, consistent with its twofold greater effective population size. B also correlates with both crossover rate and nucleotide diversity in each species. Finally, we estimate equilibrium GC content from both divergence and polymorphism data, which indicates that B has been increasing in both species, and provide support for population expansion explaining a large proportion of this increase in the zebra finch. Significance statementUnderstanding the forces that change the nucleotide base composition of genomes is central to understanding their evolution. One such force is GC biased gene conversion, a process that during recombination converts some heterozygous base positions to homozygous. This process is more likely to convert adenine and thymine bases to guanine and cytosine bases than the other way around, hence is GC biased. This increases the frequency of GC alleles in a way similar to positive selection. This process has largely been studied within protein coding regions, and not often compared between species. We measure its strength in the non-coding areas of the genomes of two bird species, showing it to be stronger in the species with the larger population size.