(1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucl. Acids Res., 25 (17), 3389-3402.
|Author||Altschul, S., Madden, T., Schaffer, A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.|
|Reference||[www.ncbi.nlm.nih.gov/pubmed/9254694 PubMed 9254694]|
|Short description||Alignment Heuristic|
|Method||Iterative & PSSM|
PSI-BLAST is one of the heuristics to calculate alignments. First, a list of all closely related proteins is created. With this result a position specific scoring matrix is calculated. This matrix is used for the next search against the database and a new PSSM is calculated. These iterations are repeated till a certain threshold or the specified number of iterations is reached.
The world of sequences can be regarded as some islands (sequence families) and the sea (sequence-empty space) between them. PSI-BLAST allows to explore such an island by extending the search space. In general this model is not fitting nature completely. The islands are less defined and there are "bridges" between them. In the early age of PSI-BLAST it was easy with too much iterations to cross the border of one sequence family and get onto a new one. Therefore iterations were often limited to three or five. But the statistics were improved in later versions of PSI-BLAST and nowadays PSI-BLAST is not needed to be restricted to a certain number of iterations.