A Graph Theoretic Algorithm for Removing Redundant Protein Sequences

Abstract

Many biological sequence databases have redundant sequences which are not helpful to statistical analysis and re- quire more computational time and resources to process. This lead us to design a new and fast program to generate a non- redundant sequence set. A graph theoretic algorithm was de- signed to process BLAST output and remove redundant pro- teins from a protein sequence database. We have developed a program named BlastCuller which can be used to generate a non-redundant protein database. BlastCuller is a flexible pro- gram with a parameter of sequence similarity cutoff, which can be a decimal from 0.0 to 1.0. This program can be down- loaded from http://pcal.biosino.org/BlastCuller.html.