Discovering Novel Protein-Coding Genes and Long Non-Coding RNAs in Bos taurus

Efficient high-throughput sequencing technologies allow us to significantly improve annotation of the bovine genome. We are using RNAseq data to re-annotate the Bos taurus UMD3.1 genome assembly and to create a reference transcriptome that includes transcripts that are not found in the current assembly. We have computed transcript assemblies for 91 RNAseq datasets, representing 95 tissues, using six different methods, which include a combination of reference-based, de novo and hybrid RNAseq assembly. Our different assembly pipelines use TopHat, STAR, Cufflinks, StringTie, Trinity and PASA. We have identified novel protein-coding genes and long-non-coding RNA after running Transdecoder to identify open reading frames (ORF) and IntersectBed to filter out transcripts overlapping Ensembl and RefSeq genes. For transcripts without ORF (candidate long non-coding RNA) we used additional filters to eliminate those with coding potential (CPAT, CPC and Blast search vs. NCBI NR). We have merged transcript assemblies from the 546 transcript assembly sets into gene loci based on overlapping genomic coordinates. The novel gene loci are available for searching by tissue and viewing in JBrowse and all of the transcript assemblies are available for BLAST search at the Bovine Genome Database (http://BovineGenome.org).