MAP: Searching Large Genome Databases

Speaker: CHENG Lok Lam

Abstract

A number of biological applications requrie comparison of large gnome sequences. Current techniques, such as BLAST, suffer from both disk I/O and computational cost because of extensive memory requriements and large candidate sets. Recently, the MAP algorithm is proposed by Tamer Kahveci and Ambj Singh to improve the speed of BLAST. MAP uses the MRS index structure to reduce both disk I/O and computational cost. The experimental results show that MAP aligns genome sequences up to 97 times faster than BLAST. In this talk, I will introduce the BLAST, MRS index structure and the MAP algorithm.