High performance computing of a fusion gene detection pipeline on the K computer

Recently developed high-throughput sequencers can generate a huge amount of omics data, and TOP500-class supercomputers are required to analyze such large datasets. However, these supercomputers do not usually support grid engines, which are commonly used in bioinformatics, making it necessary to parallelize the software. Parallelization and optimization require domain and specific knowledge, which poses a challenge for most bioinformaticians. We here propose a simple methodology for the parallelization of pipeline software. To demonstrate the efficacy of the methodology, we employed the Genomon-fusion as a sample software pipeline and ported it onto the K computer by applying our method. Simultaneous analysis of a massive amount of samples was performed using the K computer in a very short period of time.