Authors:
Álvaro Nogueira da Cruz
1
;
Vitoria Gomes
1
;
Matheus Andrade
1
;
Anderson Amorim
1
;
Carlos Valêncio
1
;
Gilberto Vaughan
2
and
Geraldo Zafalon
1
Affiliations:
1
Department of Computer Science and Statistics, Universidade Estadual Paulista (UNESP), Rua Cristóvão Colombo, 2265, Jardim Nazareth, São José do Rio Preto - SP, 15054-000, Brazil
;
2
Division of Viral Hepatitis, Centers for Diseases Control and Prevention, Atlanta, 30329, U.S.A.
Keyword(s):
Single Nucleotide Polymorphism, Next Generation Sequencing, Graphics Processing Unit, Parallel Processing, Bioinformatics.
Abstract:
In the context of bioinformatics one of the most important problems to be solved is the search for simple nu- cleotide polymorphism (SNP). When we perform the analysis of the files from the next generation sequencing (NGS) the search task for SNPs becomes more prohibitive due to the millions of sequences present on them. CPU multithreaded approaches are not enough when millions of sequences as considered. Then, the use of graphics processing units (GPUs) is a better alternative, because it can operate with hundreds of arithmetic logic units while CPU with no more than tens. Thus, in this work we developed a method to detect SNPs using a mask approach under GPU architecture. In the tests, a speedup of up to 5175.86 was obtained when com- pared to the multithreaded CPU approach, evaluating from 100,000 to 800,000 sequences using five masks to detect the occurrence of SNPs.