Rates of Evolution Differ Between Cell Types Identified by Single-Cell RNAseq in Threespine Stickleback

Rates of evolutionary change vary by gene. While some broad gene categories are highly conserved with little divergence over time, others undergo continuous selection pressure and are highly divergent among populations or species. But such comparative evolutionary studies rely on gene ontology categories that are rarely validated in the study's focal species. An alternative is to determine species-specific gene functions using single-cell RNA sequencing (scRNAseq) that identifies cell types and the characteristic genes that each expresses. Here, we combine scRNAseq with evolutionary genomics to understand whether certain cell types exhibit faster evolutionary divergence (using their characteristic genes) than other types of cells (using intestine, head kidney, liver and gill tissues from an emerging model organism, the threespine stickleback). Merging scRNAseq with population genomic data, we show that cell types differ in the rate at which their characteristic genes evolve, as measured by allele frequency divergence among many populations (F_ST) and the longer-term effect of selection driving sequence differences between species (dN/dS ratios). Neutrophils, B cells and fibroblasts exhibit elevated F_ST at characteristic genes, while eosinophils in the intestine and thrombocytes in the head kidney exhibit lower F_ST than the average for 1000 random genes (dN/dS showed similar results). Genes' positions within the scRNAseq coexpression network (measured by graph theory metrics of centrality) also differed between cell types and there was a weak tendency for genes with higher closeness centrality to have higher F_ST. These results highlight the value of merging scRNAseq technology with evolutionary population genomic data and reveal that genes which define immune cell types exhibit especially rapid evolution.