Annotating the Bovine Genome with Single-Molecule Transcript Sequencing

Bovine genome annotation with RNAseq has been challenging due to the complexity of alternative splicing, potential for read mapping errors, high repeat content of the genome and expressed sequences, and the need to assemble full-length transcripts from short reads. PacBio Iso-Seq technology for generating long-read transcript sequences helps us overcome several of these issues. Advantages of Iso-Seq are that long read lengths allow us to map transcript sequences that have repeats and to avoid predicting only partial genes. We report our approach and results of annotating the Bos taurus genome using Iso-Seq data generated from seven tissues (adipose, cerebral cortex, liver, lung, muscle, testis and thalamus). We used the long reads to update both the Ensembl and RefSeq gene sets. Changes to the gene sets included the addition of isoforms, extension of coding regions, UTR extensions, and correction of predicted genes that needed to be merged. In addition to updating existing gene sets, we identified and characterized novel gene loci.