Blast Algorithm
Blast uses pre-indexed databases
- The location of every ‘word’ of each database entry is remembered
To perform a query
- Break the query sequence into ‘words’
- Scan the database for these ‘words’
- (For each ‘word’ don’t just use the ‘word’, use all similar ‘words’)
- Where two non-overlapping ‘words’, within a certain distance of each other in the query, are matched against a database entry and with no gap required then this region of the two sequences is called a segment pair
- The segment pair is extended until the score drops by X below its maximum value
- The score for each database entry is calculated in this way
- Database entries with statistically significant scores are reported