nBLAST-JC Deployment on High-Performance Clusters

This research article presents nBLAST-JC, an accelerated version of the nucleotide BLAST algorithm designed for deployment on high-performance clusters (HPC). The objective of the study is to develop a scalable algorithm that can process multiple query sequences concurrently using the JCuda framework on the Hadoop platform. The paper details the implementation of nBLAST-JC, which utilizes the Hadoop framework to manage large volumes of molecular data efficiently. Performance testing of the proposed algorithm shows a speed-up ranging from 7.1 to 9 times compared to HS-BLASN and 1.8 to 2.3 times compared to G-BLASTN in 'blastn' mode. The article outlines the fundamental characteristics of the Hadoop framework that contribute to the scalability of nBLAST-JC for large databases. Additionally, the research discusses the integration of CPU cores with GPU resources to enhance performance further. The findings indicate that nBLAST-JC is a significant advancement in bioinformatics tools for sequence alignment.