Podium: Improving Annotated Gene Structure in the Horse with a Publically Available Transcriptome of Six Tissues

The current annotation of the horse genome is inadequate for successful genome to phenome analyses.  There is a paucity of evidence codes, and many inconsistencies exist between publicly available annotations from the NCBI and Ensembl databases. Additionally, the lack of tissue-specific transcription profiles creates an obstacle for modern functional genomics and system biology approaches. The horse genome also contains confounding frameshift errors that add another barrier for proper gene prediction. In this study, we used Tophat and Cufflinks to produce reference-guided assemblies for 69 samples with ~1.5 billion Illumina sequencing fragments. We merged the assemblies into six tissue-specific transcription profiles for cerebellum, brainstem, spinal cord, retina, muscle and skin. The final merger of all assemblies overlaps with 63% and 73% of NCBI and Ensembl loci, respectively, capturing about 72% and 81% of their coding bases. Comparing our assembly to the most recent transcriptome annotation shows ~85% overlapping loci. In addition, at least 40% of our annotated loci represent novel transcripts. As part of this work, we present a new application of digital normalization to allow for integrative analyses of RNAseq, which captures very low abundance transcripts. This enabled us to extend UTRs of our gene models and identify normally low-level transcribed ncRNA. Finally, we built a lightweight variant analysis pipeline using the GATK toolkit to detect and fix genome-sequencing errors affecting open reading frames. All of our transcriptome assemblies, a liftover of NCBI annotation and the corrected version of the horse genome are readily available as UCSC track hubs.