Population Genomic Inferences from Sparse High-Throughput Sequencing of Two Populations of Drosophila melanogaster

Short-read sequencing techniques provide the opportunity to capture genome-wide sequence data in a single experiment. A current challenge is to identify questions that shallow-depth genomic data can address successfully and to develop corresponding analytical methods that are statistically sound. Here, we apply the Roche/454 platform to survey natural variation in strains of Drosophila melanogaster from an African (n = 3) and a North American (n = 6) population. Reads were aligned to the reference D. melanogaster genomic assembly, single nucleotide polymorphisms were identified, and nucleotide variation was quantified genome wide. Simulations and empirical results suggest that nucleotide diversity can be accurately estimated from sparse data with as little as 0.2x coverage per line. The unbiased genomic sampling provided by random short-read sequencing also allows insight into distributions of transposable elements and copy number polymorphisms found within populations and demonstrates that short-read sequencing methods provide an efficient means to quantify variation in genome organization and content. Continued development of methods for statistical inference of shallow-depth genome-wide sequencing data will allow such sparse, partial data sets to become the norm in the emerging field of population genomics.