NCBI will be updating the human genome RefSeq annotation more frequently to incorporate improvements made to genes and transcripts by RefSeq curation experts. Faster updates will allow us to include the latest datasets.
In the past, we’ve produced a full re-annotation of the human genome about once a year. The last full annotation, Homo sapiens Annotation Release 109, was in March 2018. A full annotation is produced by two main processes:
- Placement of known RefSeq transcripts on the genome
- Prediction of new gene and transcript models by Gnomon
Going forward, we’ve added an update process that will incorporate the newest curated RefSeq transcripts, functional element annotations, and other improvements, but leave the model dataset the same except where they’ve been superseded by curated RefSeqs or other review. This simplified process will allow us to update the genome annotation at higher frequency for organisms where we have extensive ongoing curation efforts. We plan to update the annotation on the human reference assembly every two months. Annotation updates will be named with a date suffix, like “Annotation Release 109.20190320”.
The updated annotation will be visible on the RefSeq chromosomes and other assembly sequences in Entrez Nucleotide and NCBI Gene, and it will be available for download. It will also become the default NCBI annotation track in the Genome Data Viewer, with the original full annotation available for selection in the Genes track set.
We’ll post another announcement on NCBI Insights when the first updated annotation of GRCh38.p13 is ready to help drive your research.
5 thoughts on “Human genome annotation will be updated every 2 months”
Good news but where can I download the latest release?
Hi there! Try the Human Genome Resources page: https://www.ncbi.nlm.nih.gov/projects/genome/guide/human/
Send us an email at firstname.lastname@example.org if you need further help.
Dear NCBI, great news indeed. Could you comment on the fact that annotations can live both on the primary assembly and patches; are those synchronised in any way? Or could we imagine that the annotations (transcript wise) on the patches are more accurate than those on the primary?
Thanks for your help,