Functional convergence of genomic and transcriptomic architecture underlies schooling behaviour in a live-bearing fish

Abstract

The organization and coordination of fish schools provide a valuable model to investigate the genetic architecture of affiliative behaviours and dissect the mechanisms underlying social behaviours and personalities. Here we used replicate guppy selection lines that vary in schooling propensity and combine quantitative genetics with genomic and transcriptomic analyses to investigate the genetic basis of sociability phenotypes. We show that consistent with findings in collective motion patterns, experimental evolution of schooling propensity increased the sociability of female, but not male, guppies when swimming with unfamiliar conspecifics. This finding highlights a relevant link between coordinated motion and sociability for species forming fission-fusion societies in which both group size and the type of social interactions are dynamic across space and time. We further show that alignment and attraction, the two major traits forming the sociability personality axis in this species, showed heritability estimates at the upper end of the range previously described for social behaviours, with important variation across sexes. The results from both Pool-seq and RNA-seq data indicated that genes involved in neuron migration and synaptic function were instrumental in the evolution of sociability, highlighting a crucial role of glutamatergic synaptic function and calcium-dependent signalling processes in the evolution of schooling.

Fig. 1. Heritability of sociability in guppies.

a, Female, but not male, guppies from polarization-selected lines (n = 763, orange) presented higher alignment to the group direction (left) and shorter distances to their nearest neighbour (higher alignment; right) than guppies from control lines (n = 724, grey) when swimming with unfamiliar same-sex conspecifics (see Supplementary Tables 1 and 2). For all boxplots, horizontal lines indicate medians, boxes indicate the interquartile range and whiskers indicate all points within 1.5 times the interquartile range. Boxes in top left position of each facet indicate Tukey adjusted P values for multiple contrasts (P < 0.05 in bold) for statistical contrasts by sex in an LMM comparing alignment and attraction between selection line treatments (see Supplementary Tables 1 and 2). b, Animal models using same-sex pedigrees and full pedigrees with alignment and attraction (nearest neighbour distance) phenotypes in 195 families of polarization-selected and control guppy lines indicated a moderate heritability in female guppies for both biologically relevant aspects of sociability measured, alignment (left) and attraction (right). In males, we found moderate heritability in attraction, but CIs in alignment estimates overlapped with 0, suggesting low heritability of this sociability aspect. Our full-pedigree animal models provided large CIs for male–female correlations in sociability, with estimates overlapping 0 in alignment, but a positive correlation between sexes in attraction (see Supplementary Tables 4 and 5). Red diamonds indicate mean heritability values with 95% CIs. Source data

Fig. 2. Genetic basis of sociability in the guppy.

a, Manhattan plot of −log₁₀(P) values across linkage groups (LG) in the guppy genome resulting from a two-sided Fisher’s exact test comparing allele frequency differences between high- and low-sociability female guppies. We merged pooled DNA sequences of three independent replicates and found 819 SNPs to be significantly different (above genome-wide threshold highlighted in red), while a highly stringent analyses of consistent allele frequency differences across our three independent replicates (CMH test; see Methods) identified 13 SNPs (5 of them within genes) associated with sociability in the species (gene names and SNP location in the genome highlighted in orange). SNPs with −log₁₀(P) < 2 are omitted. b, Clustering of statistically significant overrepresented GO annotations for biological processes associated with differences between high and low sociability in guppies. Point size and colour provide information on fold enrichment value from the statistical overrepresentation test performed in PANTHER (see Methods). Terms with fold enrichment lower than 8 are represented but not described in text. Axes have no intrinsic meaning and are based on multidimensional scaling which clustered terms based on semantic similarities. Source data

Fig. 3. Neurogenomic response of schooling in guppies.

a–c, Hierarchical clustering and relative expression levels for all differentially expressed genes between Alone and Group treatments in the optic tectum (a), the telencephalon (b) and the midbrain (c). Differentially expressed genes were identified separately in polarization and control line samples. Clustering, based on Euclidian distance, represents transcriptional similarity across all samples. Venn diagrams summarize the total number of DE genes and DC gene pairs in each tissue for polarization-selected and control lines. Source data

Fig. 4. Functional characterization of genes of interest across experiments.

Visualization of functional overlap based on GO annotations between genes of interest highlighted in strongly differentiated experimental setups evaluating social interactions of female guppies following experimental evolution for higher polarization: (1) genomic analyses of DNA comparing Pool-seq of high- and low-sociability female guppies (left column); (2) transcriptomic analyses evaluating differentially expressed genes in key brain regions of polarization-selected lines of female guppies exposed to two different social contexts: swimming alone or with a group of conspecifics (TEL, telencephalon; MBR, midbrain; OT, optic tectum; columns 2–4). Shades of green indicate fold enrichment from our statistical overrepresentation tests performed to gene lists obtained from each experiment (see Supplementary Dataset). Source data