Enhanced Access to
Taxonomy Database

New Human-Mouse
Homology Map

Catch the Gene
Expression Omnibus

A Pair of Pathogens
Added to GenBank

Protein Molecular
Weight Field
Now in Entrez

OMIM In Entrez:
New Searching Power

Web Server
Software Available
for BLAST

Recent Publications

News Briefs

BLAST Lab

PSI-BLAST 2.1
Offers Composition-
Based Statistics

Slight Address Change
for NCBI FTP Server

Masthead

PSI-BLAST now permits calculated E-values to take into account the amino acid composition of the individual database sequences involved in reported alignments. Such composition-based statistical analysis improves E-value accuracy, thereby reducing the number of false positive results.

The improved statistics are achieved with a scaling procedure^1,2 that employs a slightly different scoring system for each database sequence. As a result, raw BLAST alignment scores will not correspond precisely to those implied by any standard substitution matrix. Furthermore, identical alignments can receive different scores, depending on the compositions of the sequences they involve. The improved statistics are now used by default for all rounds of searching on the Web version of PSI-BLAST, but are not used by Basic or Advanced BLAST. Therefore, if one uses default settings, the results of the first round of PSI-BLAST will be different from those obtained using the same query with Basic or Advanced BLAST.

PSI-BLAST 2.1 is currently available only on the Web, at www.ncbi.nlm.nih.gov/blast/psiblast.cgi. It will be incorporated into the standalone binaries and NCBI toolkit in the near future.
—SM

Notes
1. Altschul, SF, et al. Nucleic Acids Res 25:3389–402, 1997.
2. Schäffer, AA, et al. Bioinformatics 15:1000–11, 1999.

---