Ferret assembly and gene annotation

Assembly

The ferret (Mustela putorius furo) genome was produced by the Broad Institute. Illumina technology was used to produced this high quality draft. The whole genome shotgun data was assembled with Allpaths-LG. It is composed of 117441 contigs with an N50 value of 44.8kb and 7741 scaffolds with an N50 value of 9.3Mb.

Gene annotation

The gene set was built using a mixed approach. First the Ensembl pipeline was used to generate 264,684 models from orthologous vertebrate proteins from UniprotKB with a protein existence level of 1 or 2. Then, due to the lack of species specific sequences and the availability of RNA-Seq for ferret, we used 1.1G paired-end reads sequenced by the Broad Institute. The RNA-Seq data contains tissues from a male ferret, from a female ferret, from an influenza infected ferret and from different developmental stage. We pooled the tissues to avoid creating too many fragmented models. Using the RNA-Seq pipeline, we created 22,299 models from the pooled set. When a pooled model was missing but we had a consensus with the tissues models, the consensus model was added to the pooled set, which brought the number of RNA-Seq models up to 24,837. By combining the orthologous set, the RNA-Seq set and our ncRNA pipeline we built the final gene set: 19910 protein coding gene models, 270 pseudogenes, 17 retrotransposed and 3,614 non coding RNA.

RNASeq data set

In addition to the main set, we have predicted gene models for each tissue type using the RNA-Seq pipeline. We did a BLASTp of these models against UniProt proteins of protein existence level 1 and 2 in order to confirm the open reading frame (ORF). The best BLAST hit is displayed as a transcript supporting evidence.

The tissue-specific sets of transcript models built using our RNAseq pipeline are as follows:

• Detailed information on genebuild (PDF)

More information

General information about this species can be found in Wikipedia.