Bovine Genome Database: Working Towards an Improved Bovine Gene Set Using Community-provided RNA-Seq Data

Efficient next generation sequencing technologies have made it possible for individual research labs, to generate an abundance of transcriptome data focused on specific tissues and conditions. Incorporating these data into gene prediction pipelines will identify novel genes and alternatively spliced transcripts.  Bovine Genome Database (BGD) aims to continually serve as a resource for data generated by the community.  To this end we are working on a new gene set based on community-provided RNA-Seq data.  Using both reference-based and de novo transcriptome assembly we are implementing a variety of strategies to analyze and improve the gene set.  One such strategy is computational normalization, which was previously shown to reduce computational requirements, reduce sampling variation, and decrease the errors contained within reads.  Understanding how adjusted data parameters reduce the incidence of artificial gene splits and merges will lead to more effective gene prediction and genome annotation strategies. Additional issues associated with community-provided data must be addressed as well.  For example, some RNA-Seq libraries we acquire from the community lack crucial information such as insert size.  We continue to develop tools to overcome these issues as they are identified. We plan to optimize our pipeline using the Bos Taurus UMD3 assembly, and we will re-apply it once an improved assembly becomes available.