Molecular Ecology and Evolution of a Desert Rodent: Testing Predictions in Wild Nonmodel Species

Among the most fundamental pursuits in the realms of molecular ecology and evolution are understanding how geography and ecology drive genetic isolation, and the dissection of the genetic architecture of fitness-related traits. Since the advent of Next Generation Sequencing (NGS), these research avenues can be explored in nonmodel species, broadening the understanding of the origin and maintenance of genetic diversity in natural settings. Wild Eastern spiny mouse, Acomys dimidiatus, populations inhabit neighbouring desert wadis in South Sinai, Egypt, but parasite diversity and load appear to be different between populations. In this thesis, a suite of genetic markers, both microsatellites and potentially functional single nucleotide polymorphisms (SNPs), was developed. An mRNA-Seq experiment was described, and evidence for historical selection was examined to test the prediction that immune genes are more likely to be under positive selection than nonimmune genes. Microsatellites and potentially functional SNPs, half of which were immune non-Major Histocompatibility Complex (MHC), were typed in 363 and 254 mice respectively to elucidate population genetic structure and diversity of three cohorts of mice sampled from four wadis every three years. The influence of local conditions, as well as geographic and ecological distances, estimated from remote sensing data, on genetic isolation was investigated. Associations between individual SNP genotypes, as well as some other genetic predictors, and parasite burden were examined to test three predictions; (1) individual immune non-MHC but not control SNPs were under contemporary selection, and associated with variation in parasite burden, (2) the amount of phenotypic variance of parasite burden explained by all or immune SNPs was greater than that explained by nonimmune/control SNPs alone, and (3) Heterozygosity-Fitness Correlation (HFCs), between genomewide heterozygosity, estimated from all genetic markers, and parasite burden existed. Immune genes were not more likely to be under positive selection than nonimmune genes. Genetic diversity, inferred from 18 microsatellites and 54 SNPs, 30 of which were immune, was relatively high. A subtle but stable amount of population structure was inferred by microsatellites. However, structure was not detected by SNPs, whether using all SNPs, immune SNPs or nonimmune SNPs. Geographic and ecological distances, but not local conditions explained a fraction of genetic variability (R2= 0.078, P<0.001) but had contrasting effects; genetically similar wadis were linked by longer ecological distances but shorter geographic distances. Contemporary selection was not more likely to act on immune than nonimmune SNPs, and two SNPs, one of which was immune related, were associated with parasite burden. Contrary to the prediction, control SNPs explained more phenotypic variance of parasite burden than all or immune SNPs. No evidence for HFCs was detected between any of the heterozygosity measures and parasite burden.