Premium
Improving the sensitivity of the sequence profile method
Author(s) -
Lüthy Roland,
Xenarios Ioannis,
Bucher Philipp
Publication year - 1994
Publication title -
protein science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.353
H-Index - 175
eISSN - 1469-896X
pISSN - 0961-8368
DOI - 10.1002/pro.5560030118
Subject(s) - sensitivity (control systems) , sequence (biology) , computational biology , chemistry , biology , biochemistry , engineering , electronic engineering
The sequence profile method (Gribskov M, McLachlan AD, Eisenberg D, 1987, Proc Natl Acad Sci USA 84 :4355‐4358) is a powerful tool to detect distant relationships between amino acid sequences. A profile is a table of posi‐tion‐specific scores and gap penalties, providing a generalized description of a protein motif, which can be used for sequence alignments and database searches instead of an individual sequence. A sequence profile is derived from a multiple sequence alignment. We have found 2 ways to improve the sensitivity of sequence profiles: (1) Sequence weights: Usage of individual weights for each sequence avoids bias toward closely related sequences. These weights are automatically assigned based on the distance of the sequences using a published procedure (Sib‐bald PR, Argos P, 1990, J Mol Biol 216 :813‐818). (2) Amino acid substitution table: In addition to the alignment, the construction of a profile also needs an amino acid substitution table. We have found that in some cases a new table, the BLOSUM45 table (Henikoff S, Henikoff JG, 1992, Proc Natl Acud Sci USA 89 :10915‐10919), is more sensitive than the original Dayhoff table or the modified Dayhoff table used in the current implementation. Profiles derived by the improved method are more sensitive and selective in a number of cases where previous methods have failed to completely separate true members from false positives.