Causes and consequences of a complex recombinational landscape in the ant Cardiocondyla obscurior

Abstract

Eusocial Hymenoptera have the highest recombination rates among all multicellular animals studied so far, but it is unclear why this is and how this affects the biology of individual species. A high-resolution linkage map for the ant Cardiocondyla obscurior corroborates genome-wide high recombination rates reported for ants (8.1 cM/Mb). However, recombination is locally suppressed in regions that are enriched with TEs, that have strong haplotype divergence, or that show signatures of epistatic selection in C. obscurior The results do not support the hypotheses that high recombination rates are linked to phenotypic plasticity or to modulating selection efficiency. Instead, genetic diversity and the frequency of structural variants correlate positively with local recombination rates, potentially compensating for the low levels of genetic variation expected in haplodiploid social Hymenoptera with low effective population size. Ultimately, the data show that recombination contributes to within-population polymorphism and to the divergence of the lineages within C. obscurior.

Figure 1.

Recombination rate variation and genome architecture in C. obscurior. (A) Circos plot showing the following inward: (SV) the location of identified structural variants (dark green), (RR) recombination rates across the genome (cM/Mb) with introgressions depicted with purple bars, (TE) TE content (“LDRs”, blue; “TE islands”, orange) with the location of TE islands shown as black bars, (Ex) exon content, and (π) the log₁₀-transformed nucleotide diversity in two natural populations for C. obscurior from Itabuna (gold) and Una (maroon). C. obscurior queen with brood (photo: L. Schrader). (B) Recombination rates between TE-poor regions (“LDRs” in blue) and TE-rich regions (“TE islands” in orange). (C) Recombination rates in introgressed regions (“LDRs introgr.” In purple) compared with the remainder of LDRs (“LDRs other” in blue). (D) Levels of nucleotide diversity (π) in nonrecombining (“LDRs nonrec.”) compared with recombining regions LDRs (“LDRs rec.”) in two populations of C. obscurior from Itabuna and Una. (****) P < 0.0001, (**) P < 0.01.

Figure 2.

Relationship between recombination rates and measures of diversity in populations of C. obscurior. (Left) Correlation between recombination rates and nucleotide diversity and Tajima's D. (Middle) Correlation between recombination rates and nucleotide diversity at synonymous (π_S) and nonsynonymous (π_NS) sites. (Right) Correlation between recombination rates and levels of genetic differentiation (F_ST) and absolute divergence (d_XY) between lineages of C. obscurior. The x-axis represents binned recombination rates, with the upper limit of each bin indicated and “TEI” referring to TE islands (orange boxplot). (ρ) Spearman's rank correlation coefficient, with significance shown by asterisks: (****) P < 0.0001, (***) P < 0.001, (**) P < 0.01. All correlations were performed using the raw data. Lines and shaded areas are linear regressions and 95% confidence intervals.

Figure 3.

Association between recombination rates and genome dynamics in C. obscurior. Correlations between recombination rates and length (A) and frequency (B) of SVs and indel polymorphism (C) in 16 individual workers. The x-axis represents binned recombination rates, with the upper limit of each bin indicated. (ρ) Spearman's rank correlation coefficient is indicated for all structural variants (SVs) and separately for deletions (DEL), duplications (DUP), and inversions (INV), with significance shown by asterisks: (****) P < 0.0001, (***) P < 0.001, (**) P < 0.01. All correlations were performed using the raw data. Red lines and shaded areas are linear regressions and 95% confidence intervals.