DocumentCode
191011
Title
SAIS-OPT: On the characterization and optimization of the SA-IS algorithm for suffix array construction
Author
Timoshevskaya, Nataliya ; Wu-Chun Feng
Author_Institution
Dept. of Comput. Sci., Virginia Tech, Blacksburg, VA, USA
fYear
2014
fDate
2-4 June 2014
Firstpage
1
Lastpage
6
Abstract
The suffix array and Burrows-Wheeler Transform are critical index structures in next generation sequence analysis. The construction of such index structures for mammalian-sized genomes can take thousands of seconds (i.e. tens of minutes). Its construction is complicated by computational overheads that coming from irregular or complex memory-access patterns. This paper rigorously characterizes the execution profile of the SA-IS algorithm in order to guide its optimization. The resulting optimized SA-IS, which we refer to as sais-opt, outperforms previous implementations of SA-IS as well as “best in practice” algorithms, when applied to large DNA strings.
Keywords
DNA; bioinformatics; genomics; optimisation; Burrows-Wheeler transform; DNA string; SA-IS algorithm characterization; SA-IS algorithm optimization; SAIS-OPT; complex memory-access patterns; irregular memory-access patterns; mammalian-sized genomes; next generation sequence analysis; suffix array construction; Arrays; Classification algorithms; Genomics; Indexes; Optimization; Random access memory; Sorting; Burrows-Wheeler Transform; irregular memory access; suffix array;
fLanguage
English
Publisher
ieee
Conference_Titel
Computational Advances in Bio and Medical Sciences (ICCABS), 2014 IEEE 4th International Conference on
Conference_Location
Miami, FL
Print_ISBN
978-1-4799-5786-6
Type
conf
DOI
10.1109/ICCABS.2014.6863917
Filename
6863917
Link To Document