The smallest patches in NYC often contain the highest population densities of white-footed mice, most likely due to ecological release and obstacles to dispersal. Consistently elevated population density in urban patches compared to surrounding rural populations is predicted to result in density-dependent selective pressures on traits related to immunology, intraspecific competition, and male-male competition for mating opportunities, among others. White-footed mouse populations in NYC exhibit high levels of heterozygosity and allelic diversity at neutral loci within populations, but genetic differentiation and low migration rates between populations. This genetic structure contrasts with weak differentiation reported for Peromyscus spp. at regional scales, or even between populations isolated on different islands for thousands of generations. High genetic diversity within and low to nonexistent migration between most NYC populations suggests that selection can operate efficiently within these geographically isolated populations, either on standing genetic variation or de novo mutations. In this study we take steps to develop P. leucopus as a genomic model for adaptive change in urban environments. Pooling mRNA from multiple individuals is an effective approach to AbMole Folic acid transcriptome sequencing that avoids the prohibitive cost of sequencing individual genomes. While pooling results in the loss of genetic information from individuals, the ability to identify SNPs in a population increases due to the inclusion of multiple individuals in the pool. By analyzing SNPs within thousands of transcripts, it is feasible to identify candidate genes underlying rapid divergence of populations in novel environments. Certain statistical approaches, such as the ratio between nonsynonymous and synonymous substitutions, can be applied to pooled transcriptome data to identify potential signatures of selection between isolated populations. If positive selection is acting on a codon, then nonsynonymous AbMole Enoxacin hydrate mutations should be more common than under neutral expectations. Here, we describe the results of de novo transcriptome sequencing, annotation, SNP discovery, and outlier scans for selection among urban and rural white-footed mouse populations. We used the 454 GS FLX+ system to sequence cDNA libraries generated from pooled mRNA samples from multiple tissues and populations. Several de novo transcriptome assembly programs were used and the contribution of specific tissue types to the transcriptome assembly was examined. We then identified coding region SNPs between urban and rural populations, and scanned this dataset for signatures of positive selection using pS between populations and McDonald-Kreitman tests between multiple species. We report several candidate genes potentially experiencing directional selection in urban environments, and provide annotated transcriptome datasets for future evolutionary studies of an emerging model system. Despite the substantially fewer megabases per run generated by 454 FLX+ compared to Illumina or SOLiD sequencing, we still ran into computational limitations during assembly when using options for cDNA sequence. Similar to Cahais et al., we had the most success after compressing the raw reads into a smaller number of partially assembled sequences using a genome assembler followed by another assembly method better suited for transcriptome data. While the CAP3 assembly produced more contigs, the Newbler v. 
2.5.3 transcriptome assembly performed better based on assessments useful for downstream population genomic analyses.