Population and metabolic genomics of five geographically dispersed fully-sequenced population samples of Drosophila melanogaster. Andrew G. Clark¹, J. Roman Arguello¹, Margarida Cardoso Moreira¹, Jian Lu¹, Cornelia J. Scheitz¹, Anthony J. Greenberg¹, Sean R. Hackett^1,2, Julien F. Ayroles^1,3, Srikanth Gottipati¹, Lawrence G. Harshman⁴, Jennifer K. Grenier¹. 1) Dept Molec Biol & Gen, Cornell Univ, Ithaca, NY; 2) Grad Program Quant and Comp Biology, Princeton Univ, Princeton, NJ; 3) Dept of OEB, Harvard Univ, Cambridge MA; 4) School of Biol Sciences, Univ Nebraska, Lincoln, NE.

   In order to assess the role of geographic subdivision on the genetics of complex traits in Drosophila melanogaster, 92 inbred lines were established by sib mating isofemale stocks from Beijing, Ithaca, Netherlands, Tasmania and Zimbabwe. Genomic DNA from these lines was sequenced to approximately 12x coverage, aligned to the reference sequence and SNPs were called with the GATK pipeline. SNP genotype calls were validated with double-digest GBS and deep (100x) whole-genome sequencing of a subsample. Small indels, copy number variants (including novel genes) and inversions were also discovered and validated. Many metabolic phenotypes have been quantified in nested and well replicated designs of the inbred lines and a partial F1 diallel, including kinetics of 22 enzymes, whole-transcriptome microarrays, respirometry, flight performance, metabolites, and lipidomics. Hierarchical and generalized linear models find extensive geographic differences in genetic architecture of key metabolic traits as well as widespread genotype x environment interaction. The sequence data were of sufficient quality to provide useful inference of demographic parameters including effective size in each sample, founding times and migration rates (inferred by approximate Bayesian computation). A surprising result is the degree of both phenotypic and genome sequence differentiation of the Beijing lines, and the strength of evidence for back migration into Africa. There are large differences in distributions of transposon abundance, and even the lipidomics data show a strong signature of population differentiation. These lines augment the DGRP reference panel by adding a geographic dimension to genomic variation and adaptation.