We will continue to target manual curation to improve structural and functional information for human and other vertebrate genomes. Our conservative manual curation approach ensures the continued high quality and reliability of the human, mouse, and other ‘known’ RefSeq records which serves the needs of those who need a well-supported definition of alternate exons (fewer false positives). The addition of RNA-Seq data to our annotation pipeline significantly increased our annotation of alternate splice variants as model RefSeqs to serve the needs of those who want a more comprehensive, but still well-supported, definition of the exome (fewer false negatives). While both known and model RefSeqs report the support evidence on the sequence record, they use distinct approaches to do so. Future efforts will be directed toward harmonizing evidence reporting for both ‘known’ and ‘model’ RefSeqs so that users can more easily identify this information. We will also be adding a new data type to the human and mouse RefSeq collection in the near future to represent experimentally reported regulatory and functional elements with known (or reasonably inferred) functional consequences.
