|
MATCHER application details page
Finds the best local alignments between two sequences
matcher compares two sequences looking for local sequence similaritiesusing a rigorous algorithm. matcher is based on Bill Pearson's 'lalign' application, version 2.0u4Feb. 1996 Lalign uses code developed by X. Huang and W. Miller (Adv. Appl. Math. (1991) 12:337-357) for the "sim" program, which is a linear-spaceversion of an algorithm described by M. S. Waterman and M. Eggert (J. Mol. Biol. 197:723-728). Like water, matcher is rigorous, but also very slow. Theadvantage of matcher is that it uses far less memory thanwater, so you are much less likely to run out of memory whenaligning large sequences. matcher will also report a specified number of alignments betweenthe two sequences showing the actual local alignments. (waterwill only report the single best match.) The default number ofalignments output is 1, but can be increased to (for example) the 10best alignments by using the '-alternatives 10' command-line qualifier. In some cases, for example multidomain proteins or cDNA and genomic DNAcomparisons, there may be many interesting and significant alignments.
water will give a single best rigorous local alignment. It willuse memory of the order of the product of the lengths of the sequencesto be aligned. If you wish the 'best' local alignment you should usewater. If you run out of memory or want several possible goodalignments, use matcher.
|