NCBI staff started transitioning to Accession-Based BLAST in December 2018. The new system includes of a new version of BLAST+ executables that runs against newly-created accession-based BLAST databases. In contrast to the traditional BLAST databases that operate on the GeneInfo identifiers (GIs), the new databases use accession numbers as the only type of record identifiers.
Accession-Based BLAST stems from the NCBI-wide activity of phasing out of GIs in the sequences databases. We are now maintaining an increasing number of sequence records that no longer contain GIs as their secondary identifiers. While we are unable to include such records in the traditional BLAST databases, the new system allows us to do so.
For most web BLAST users, the transition to the new system won't be obvious as the executables interact with the databases behind the scenes. Moreover, we are transitioning in stages. As of January 2019, the Nucleotide BLAST suite continues to use traditional BLAST databases. In the Protein suite (blastp), some of the databases that you see listed on the Database menu are already accession-based (for example Metagenomic Proteins). Once you select a database, the absence of the Entrez Query option will indicate that you are searching against an accession-based database.
Refer to additional articles on the topic, especially if you are a user of BLAST+ executables (commands-line tools):
- advantages of accession-based BLAST
- obtaining executables and working with accession-based BLAST
- the future of the traditional BLAST databases