Annotating Genomes, Assessing Quality, and Downloading Data

The NCBI Eukaryotic Genome Annotation Pipeline (www.ncbi.nlm.nih.gov/genome/annotation_euk/) has been used to annotate over 350 organisms, including plants and animals of agricultural importance.

The pipeline uses a modular framework for the automated execution of all annotation tasks from the fetching of genome assemblies to the loading of RefSeq-accessioned annotation products to public databases. The annotation of an assembly takes an average of 6 days depending on the size of the genome and the amount of same-species or close cross-species expression data in public databases. Proteins, ESTs, full-length cDNAs and RNA-Seq data are retrieved, aligned to the genome of interest, and used by Gnomon, an alignment and HMM based program developed at NCBI, to predict gene models. After consolidation of the Gnomon models with RefSeq-curated transcripts and filtering of low-confidence predictions, final transcripts and proteins are named based on orthology to model organisms or Blast hits to SwissProt/UniProtKb. The annotated products are available for download and distributed through various NCBI resources including the sequence databases, Gene, BLAST and the Genome Data Viewer. A web page summarizing the annotation results is provided with each annotated genome.

This talk will provide a high-level view of the annotation pipeline. We will show how we assess the quality of the annotation products that we generate, how contiguity and accuracy of genome assemblies affect the results, and how we use transcriptomic data to compensate for assembly issues in the annotated transcripts and proteins. Finally, we will illustrate how to access assemblies and annotation results in NCBI resources.

See all eukaryotes annotated by the NCBI Eukaryotic Annotation Pipeline at: http://www.ncbi.nlm.nih.gov/genome/annotation_euk/all/