Core genes can have higher recombination rates than accessory genes within global microbial populations

Values listed in the table were computed for core/accessory genome pairs for individual clusters and pairs of clusters. Effect sizes for the relative recombination rate (${\left(\gamma /\mu \right)}_{pool}$) were calculated using Cohen’s d for paired samples using the mean paired difference ($〈{\left(\gamma /\mu \right)}_p\left(acc\right)-{\left(\gamma /\mu \right)}_p\left(core\right)〉$) divided by the standard deviation of paired differences ($\mathrm{\sigma }$). Two-sided p values were calculated using the Wilcoxon signed-rank test to compare the paired distributions of relative recombination rates for the core and accessory genome within each microbial species. The null hypothesis is that the distribution of paired differences is symmetric about zero (where $X^i\left(acc\right)-X^i\left(core\right)$ is the paired difference for recombination parameter $X$ inferred for the core and accessory genome of cluster or cluster pair i). The columns ${\left(\gamma /\mu \right)}_{pool}\left(acc\right)$ and ${\left(\gamma /\mu \right)}_{pool}\left(core\right)$ correspond to the median relative recombination rate of the accessory and core genome over all clusters and pairs of clusters. Values displayed in brackets are 95% bootstrap CIs calculated by sampling the distributions with replacement 10,000 times.

Values listed in the table were computed using the pool’s mutational divergence (${\theta }_{pool}$) for core/accessory genome pairs for individual clusters and pairs of clusters. This table is identical to Supplementary file 1 except the statistics displayed here were calculated using ${\theta }_{pool}$ .

Column descriptions are as follows: ‘species’ and ‘full name’ give the abbreviated species name used in the main text and full species name, respectively, ‘major clusters’ gives the number of sequence clusters analyzed, ‘total strains aligned’ gives the number of consensus genomes (or whole genome sequences) which were made by aligning raw reads to the reference genome, ‘total strains in major clusters’ gives the number of strains included in the major clusters (and therefore included in the analysis of recombination parameters), and ‘min strains per cluster’ gives the minimum number of strains a cluster had to have to be designated as a major cluster. The minimum cluster size was lowered for smaller strain collections as follows (where s is the number of strains in the collection): for s > 10,000 in the collection the minimum cluster size was 100, for 10,000 > s > 5000 the minimum was 50 (S. enterica was on the border, so we took the minimum to be 50), for 5000 > s > 1000 the minimum was 25, and for s < 1000 the minimum was 10. The last four columns give the average synonymous diversity ($d_s$) within clusters and between clusters (shown as mean ± st. dev.). ‘Acc’ stands for accessory.

Two-sided p values were calculated using the Friedman test to compare the matched distributions of the normalized recombination rate (${\displaystyle \gamma /{\mu }_{pool}}$) and recombinational coverage ($c_{sample}$) for all four gene frequency bins within each microbial species shown in Figure 3—figure supplement 1. Null hypothesis in this case is that the measurements have been drawn from the same distribution.

Source of data and corresponding identifiers for all sequencing data used in this study.