基于纳米计算结构上的生物序列模体发现算法  

Motif Finding Algorithm on Nanocomputing Platform

作　　者：沈一飞[1] 陈国良[1] 张强峰[1] 

机构地区：[1]中国科学技术大学计算机科学技术系高性能计算及应用省部共建重点实验室,安徽合肥230027

出　　处：《小型微型计算机系统》2007年第4期635-639,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60533020)资助.

摘　　要：模体发现对于预测基因特殊功能位点和鉴别药物作用目标等有重要的应用价值.本文介绍了一种纳米计算平台系统结构模型—CellMatrix以及在其上实现的DNA序列模体发现算法.CellMatrix是一种针对纳米计算平台提出的由同构晶格组成的可重构系统结构.这种结构既便于大规模工业生产,也很容易使得各种计算机软硬件系统在其上实现,同时这种结构又具备良好的可扩放性,是未来实现成熟纳米计算平台的一种选择.基于CellMatrix结构,本文首先在晶格结构上设计基本字符比较单元,而后在此基础上逐层构建更高层次的子序列测试单元和单条序列处理架构,从而实现了基于模式驱动的模体发现算法.最后用晶格开销数目和晶格延迟给出该算法的时空开销.Motif finding plays a key role in the field of functional sites prediction and drug design. This paper introduces a new computing architecture called Cell Matrix for nanocomputing platform and presents a pattern-driven motif finding algorithm on Cell Matrix as well. As a reconfigurable parallel architecture for nanocomputing, Cell Matrix is composed of homogenous cells that are convenient both for hardware manufacturing and software realization. Together with its scalability, this architecture is a good choice for future nanocomputers. This paper implements the pattern-driven motif finding algorithm on Cell Matrix by firstly designing the basic comparing unit and then hierarchically constructing the higher subsequence test unit and the architecture for processing one entire sequence. The complexity of algorithms in terms of cell numbers and cell delays is also studied.

关 键 词：纳米计算平台 模体发现 CELL MATRIX 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]