Sequence alignment method and system for next-generation sequencing data based on many-core platform

• Main Authors: ZHANG JINXIAO, YIN ZEKUN, LIU WEIGUO, LIU MEIYANG, ZHANG WEN
• Format: Patent
• Language: Chinese; English
• Published: 24.07.2020
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS

The invention discloses a sequence alignment method and system for next-generation sequencing data based on a many-core platform, and the method comprises the steps: adaptively adjusting the size of aread-in data block according to the calculation capability of the platform, and achieving the mutual covering of calculation and I/O; reducing memory overhead by utilizing a pre-allocation strategy and a lightweight memory allocation strategy based on a memory pool strategy; fully utilizing a vector processing unit of a processor, and adopting an instruction set for vectorizing and parallelizingthe calculation process. For a BWA-MEM algorithm, the overall performance of the method is remarkably improved, and the overall performance of a program achieves 3.62 times of the acceleration ratio;the algorithm thread expansibility is obviously improved; compared with the original algorithm, the core calculation part has the advantage that the acceleration ratio obtained under data sets with different sizes is 8.4-12.6.