冷冻电镜图像生物大分子颗粒识别方法研究  

A Review of Automatic Particle Recognition in Cryo-EM Images

作　　者：巫小蓉[1,2] 吴效明[3] 

机构地区：[1]华南理工大学计算机科学与工程学院,广州510641 [2]广东外语外贸大学信息科学技术学院,广州510420 [3]华南理工大学生物工程学院,广州510641

出　　处：《生物医学工程学杂志》2010年第5期1178-1182,共5页Journal of Biomedical Engineering

基　　金：广东省高等院校学科建设专项资金资助

摘　　要：冷冻电镜(Cryo-EM)图像生物大分子颗粒识别是Cryo-EM单颗粒三维重构非常重要的环节。为了满足三维结构分辨率越来越高的要求,需要提取更多的大分子颗粒投影。人工颗粒选择尽管有计算机辅助预选,仍然成为整个三维重构的瓶颈。Cryo-EM图像低信噪比和低对比度的特点增大了生物大分子颗粒自动识别的难度。这引起了研究者的关注,很多方法被提出。阐述模板匹配、边缘检测和图像分割方法、基于特征的方法、神经网络、基于DoG的方法和模拟退火算法在Cryo-EM图像生物大分子颗粒识别中的运用,讨论各种方法的特点,并初步探讨未来研究及改进的方向。Advances in cryo-electron microscopy（Cryo-EM） and single-particle reconstruction have led to increasingly high resolutions of macromolecular three-dimensional reconstruction.However,for keeping up the continuing improvements in resolution,it is necessary to increase the number of particles included in performing reconstructions.Manual selection of particles,even assisted by computer,is a bottleneck of single-particle reconstruction.Cryo-EM image has low signal-tonoise ratio and low contrast,which leads to difficulty in particle picking.Various approaches have been developed to address the problem of automatic particle.This paper describes the application of template-based method,edge based method, feature-based method,neural network,DoG-based and simulated annealing approach in particle picking.The characteristics of various approaches are discussed,and the future development is presented.

关 键 词：颗粒检测 生物大分子三维重构 冷冻电镜 交叉相关 模式识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]