The Ferret Genome or How RNAseq Data Helped the Annotation

With the emergence of transcriptome data generated by Next Generation Sequencing, large sets of RNA-Seq data have been made available alongside genome assemblies for many species including human, zebrafish, tilapia and ferret. These RNA-Seq data are particularly important for gene annotation when the species have few protein or cDNA sequences in public databases, as the data can inform species-specific transcript splicing and expression.

We annotated the ferret genome using our standard annotation pipeline and RNA-Seq pipeline, using data from healthy and influenza infected individuals, different tissues and different developmental stage provided by the Broad Institute.

To generate the RNA-Seq models, we aligned paired-end reads to the genome using BWA to determine transcribed regions (alignment blocks). Mate-pair information was used to group the alignment blocks into approximate transcript structures called proto-transcripts. Partially mapped raw reads were mapped a second time, this time to the proto-transcripts, using a splice-aware alignment model. Spliced reads were used to identify exon-intron boundaries and to generate transcript models. Then we filtered the models with Uniprot and added the aligned transcript models to our standard pipeline. Transcript models were created for each data set and also from the pooled set of data.

Users can view the gene models, raw bwa alignments and intron features predicted by our RNAseq pipeline alongside the ferret gene set on the Ensembl genome browser http://www.ensembl.org/Mustela_putorius_furo.