タンデム質量分析ソフトウェアCoCoozoのマルチコアCPUとGPUを用いた高速化  [in Japanese] Acceleration of Tandem Mass Spectra Analysis Software CoCoozo using Multi-core CPUs and Graphics Processing Units  [in Japanese]

Access this Article

Search this Article

Author(s)

    • 小幡 康文 Yasufumi Obata
    • 東京工業大学工学部情報工学科 Department of Computer Science, Faculty of Engineering, Tokyo Institute of Technology
    • 石田 貴士 Takashi Ishida
    • 東京工業大学大学院情報理工学研究科計算工学専攻 Department of Computer Science, Graduate School of Information Science and Engineering, Tokyo Institute of Technology
    • 夏目 徹 Tohru Natsume
    • 産業技術総合研究所創薬分子プロファイリング研究センター Molecular Profiling Research Center for Drug Discovery, National Institute of Advanced Industrial Science and Technology
    • 秋山 泰 Yutaka Akiyama
    • 東京工業大学大学院情報理工学研究科計算工学専攻 Department of Computer Science, Graduate School of Information Science and Engineering, Tokyo Institute of Technology

Abstract

生命科学や創薬などの分野において,タンパク質を同定する手法の一つに,タンデム質量分析があるが,近年,機器の発展やデータベースの増大によって,質量分析において計算機による解析が律速となりつつある.そこで本研究では,この問題に対処するために,質量分析スペクトル解析ソフトウェア CoCoozo を対象にして,高速化を目的に改良を行った.本研究では,アルゴリズムの改良に加え,マルチスレッド化,GPGPU 化を行い,プレカーサ情報完備の場合の解析について,アルゴリズムの改良だけで,CPU でも従来と比べて 8.9 倍の高速化を実現した.さらに,プレカーサ情報が欠落した場合の解析においては,12 CPU コアを用いた場合で,従来に比べて 15.9 倍,それに加えて GPU を用いた場合で,従来と比べて 18.1 倍の高速化を実現した.Tandem mass spectrometry, a method involving multiple steps of mass spectral selection, is widely used in various biological fields. In recent years, steady improvements have been made with respect to speed, and the number of protein databases available for analysis has rapidly increased. Consequently, computational analysis has become the bottleneck in tandem mass spectrometry. To overcome this problem, we attempted to improve the tandem mass spectrometry analysis software CoCoozo. To accelerate the program, we improved the algorithm and also incorporated utilization of multi-core CPU and GPGPU. As a result, when all mass spectral data files had precursor data, we achieved 8.9-fold speedups compared with the original software. In addition, in the case of no precursor data, by using a 12-core CPU and a GPU card we achieved 18.1-fold speedups compared with the original software.

Tandem mass spectrometry, a method involving multiple steps of mass spectral selection, is widely used in various biological fields. In recent years, steady improvements have been made with respect to speed, and the number of protein databases available for analysis has rapidly increased. Consequently, computational analysis has become the bottleneck in tandem mass spectrometry. To overcome this problem, we attempted to improve the tandem mass spectrometry analysis software CoCoozo. To accelerate the program, we improved the algorithm and also incorporated utilization of multi-core CPU and GPGPU. As a result, when all mass spectral data files had precursor data, we achieved 8.9-fold speedups compared with the original software. In addition, in the case of no precursor data, by using a 12-core CPU and a GPU card we achieved 18.1-fold speedups compared with the original software.

Journal

  • IPSJ SIG Notes

    IPSJ SIG Notes 2013-MPS-94(5), 1-4, 2013-07-15

    Information Processing Society of Japan (IPSJ)

Codes

  • NII Article ID (NAID)
    110009586919
  • NII NACSIS-CAT ID (NCID)
    AN10505667
  • Text Lang
    JPN
  • Article Type
    Technical Report
  • ISSN
    09196072
  • Data Source
    NII-ELS  IPSJ 
Page Top