Accelerating Gene Identification in DNA Sequences with CUDA and OpenCL

The new generation of genetic sequencing devices (YND) can generate hundreds of millions, or even billions, of sequences at a time. Applications such as Orphelia and MGC, which use conventional methods for detecting the genes in these sequences produced by YND devices, are inadequate. These applications, which do not have the capacity to process millions of sequences, take days to identify genes when the sequence file is divided into smaller groups (20K sequence). In this study, MGC application developed for metagenome gene detection, core functions are prepared for CUDA and OpenCL platforms by making appropriate data transformations. These core functions are run on the GPU and the results are discussed. With the simplification of the data structure, the gene detection process taking days is reduced to a few minutes.