Similarity Search and Multiple Sequence Alignment

There are two basic search strategies for finding a set of related sequences, namely keywords and similarity. A keyword search identifies sequences by looking through their written descriptions (i.e. the annotation section of a database file) and a similarity search looks at the sequences themselves. The keyword search is easier and seems more intuitive, but it is far from exhaustive. The main search engines for similarity search/ alignment are Entrez (NCBI) at http //www.ncbi.nhn.nih.gov/Entrez/ and SRS at http //srs.ebi.ac.uk/ 

Probe each sequence in the chosen database independently for local regions of similarity to the query sequence, using a BLAST type search but allowing gaps. 

Collect significant hits and construct a multiple alignment table between the query sequence and the significant local matches. 

Form a profile from the multiple sequence alignment. 

Reprobe the database with the profile, continuing to look only for local matches. 

---

Gracy, J. Argos, P. (1999). Automated protein sequence database classification. I. Integration of compositional similarity search, local similarity search, and multiple sequence alignment. Bioinformatics 14,164-73. 

---