The nature and distribution of non-functional alleles suggest only two independent events at the origins of Astyanax mexicanus cavefish populations

Abstract Background Several studies suggested that cavefish populations of Astyanax mexicanus settled during the Late Pleistocene. This implies that the cavefish’s most conspicuous phenotypic changes, blindness and depigmentation, and more cryptic characters important for cave life, evolved rapidly. Results Using the published genomes of 47 Astyanax cavefish from la Cueva de El Pachón, El Sótano de la Tinaja, La Cueva Chica and El Sótano de Molino, we searched for loss-of-function mutations in previously defined sets of genes, i.e., vision, circadian clock and pigmentation genes — non-functional alleles for four vision genes were identified. Then, we searched genome-wide for non-functional alleles in these four cave populations. Among 512 genes with segregating non-functional alleles, we found an enrichment in visual perception genes. Among cavefish populations, different levels of shared non-functional alleles were found. Using a subset of 12 genes for which loss-of-function mutations were found, we extend the analysis of shared pseudogenes to 11 cave populations. Using a subset of six genes for which loss-of-function mutations were found in the El Sótano del Toro population, where extensive hybridization with surface fish occurs, we found a correlation between the level of eye regression and the amount of non-functional alleles. Conclusions We can confirm that very few non-functional alleles are present in a large set of vision genes, in accordance with the recent origin of Astyanax mexicanus cavefish. Nevertheless, the genome-wide analysis indicates an enrichment of loss-of-function alleles in genes with vision-related GO-terms, suggesting that visual perception may be the function chiefly impacted by gene losses related to the shift from a surface to a cave environment. The geographic distribution of the loss-of-function alleles supports the hypothesis that cave populations from Sierra de Guatemala and Sierra de El Abra share a common origin, albeit followed by independent evolution for a long period of time. It also supports that populations from the Micos area have an independent origin. In El Sótano del Toro, the troglomorphic phenotype is maintained despite massive introgression of the surface genome.