Despite two decades of research and extensive restoration efforts since the Exxon Valdez oil spill in 1989, Pacific herring populations in Prince William Sound have not fully recovered. While the causes for the initial collapse of Prince William Sound herring are not completely clear, it is essential to identify current population structure to be able to better understand dynamics of recolonization. This information is critical for Prince William Sound restoration efforts. The challenge is that current genetic information does not allow for the increased resolution necessary to distinguish populations in the region. However, the advent of new genetic technologies allowing the sequencing of the entire expressed genome (transcriptome) of non-model species provides the opportunity for large-scale genetic marker discovery. Pyrosequencing was carried out on Pacific herring liver and gonad libraries and over 16,000 putative SNPs were identified. A subset of these SNPs were selected for validation based on quality parameters and functional annotation of transcripts. These results indicate deep sequencing of target transcriptomes is a viable method for genetic marker identification in non-model species.