Demography and adaptation promoting evolutionary transitions in a mammalian genus that diversified during the Pleistocene

Species that evolved in temperate regions during the Pleistocene experienced periods of extreme climatic transitions.  Consequent population fragmentation and dynamics had the potential to generate small, isolated populations where the influence of genetic drift would be expected to be strong.  We use comparative genomics to assess the evolutionary influence of historical demographics and natural selection through a series of transitions associated with the formation of the genus Capreolus, speciation within this genus during the Quaternary and during divergence among European roe deer (C. capreolus) populations. Our analyses were facilitated by the generation of a new high-coverage reference genome for the Siberian roe deer (Capreolus pygargus).  We find progressive reductions in effective population size (Ne), despite very large census sizes in modern C. capreolus populations and show that low Ne has impacted the C. capreolus genome, reducing diversity and increasing linkage disequilibrium.  Even so, we find evidence for natural selection shared among C. capreolus populations, including a population that has been through a severe bottleneck. During each period of transition there is evidence for selection, including at loci associated with diapause (delayed embryonic development), a phenotype restricted to this genus among the even-toed ungulates.  Together these data allow us to assess expectations for the origin and diversification of a mammalian genus during a period of extreme environmental change.

Methods
SNP datasets for European roe deer (Capreolus capreolus) samples generated with the ddRADseq protocol, in PED and MAP format (plink format).

SNPs were called using the STACKS refmap pipeline, using the Siberian roe (Capreolus pygargus) genome as reference.

The main dataset, generated with the restriction enzymes HindIII and MspI, contains samples from East Anglia (England), Ayrshire (Scotland) and Wurttemberg (Germany).

The Aurignac dataset, generated with the restriction enzymes EcoR1 and MspI, contains samples from Aurignac (France).

The commands.txt file contains all the commands used to analyse the SNP datasets.