Options for running PSI-BLAST and processing of PSI-BLAST alignments

PSI-BLAST is run for each representative sequence against the UNIREF90 database and the PSI-BLAST alignment is processed so that divergent homologs to the query are removed and a limited number of the remaining homologs are kept to save time for building an amino acid profile for the query.


     - PSI-BLAST iteration number: maximum number of iterations done by PSI-BLAST.

     - PSI-BLAST e-value inclusion threshold: only hits with an e-value less than this threshold will be included in the next iteration.

      - Identity cutoff below which distant homologs are removed: any PSI-BLAST hit with a sequence identity to the query less than this cutoff will be removed. Divergent homologs could negatively affect sequence profile quality. (default: 0.2, corresponding to 20% sequence identity)

     - Maximum number of homologs kept for each blast run: after removing divergent hits, select up to this number of homologs for constructing an amino acid sequence profile for the query. (default: 300)