Design of parallel graph-based protein sequence clustering algorithm

Abstract

Clustering protein sequences is becoming important in helping biologists analyze the large protein sequences produced by wet lab experiments. Graph based partitioning methods which is an important and stable algorithm in computer science can be used to cluster protein sequences such that each identified sub-graph can be considered as a cluster. Each cluster represents a family of protein sequences or protein sequences that shared a common attribute. Since the size of protein sequence databases increases 1.5 times yearly, a fast and efficient graph based protein sequence clustering method is much needed. We graph-based proposed a parallel approach in clustering methods by improving the performance of an existing algorithm ProtClust by using parallel methods. We presented the design of a parallel method which will be the basis of our experiments for protein sequence clustering.