Bioinformatics Institute 30 Biopolis Street, #07-01 Matrix, Singapore.
Nucleic Acids Res. 2010 Jul;38(Web Server issue):W590-4. doi: 10.1093/nar/gkq489. Epub 2010 Jun 3.
deconSTRUCT webserver offers an interface to a protein database search engine, usable for a general purpose detection of similar protein (sub)structures. Initially, it deconstructs the query structure into its secondary structure elements (SSEs) and reassembles the match to the target by requiring a (tunable) degree of similarity in the direction and sequential order of SSEs. Hierarchical organization and judicious use of the information about protein structure enables deconSTRUCT to achieve the sensitivity and specificity of the established search engines at orders of magnitude increased speed, without tying up irretrievably the substructure information in the form of a hash. In a post-processing step, a match on the level of the backbone atoms is constructed. The results presented to the user consist of the list of the matched SSEs, the transformation matrix for rigid superposition of the structures and several ways of visualization, both downloadable and implemented as a web-browser plug-in. The server is available at http://epsf.bmad.bii.a-star.edu.sg/struct_server.html.
deconSTRUCT 网络服务器提供了一个蛋白质数据库搜索引擎的接口,可用于通用的相似蛋白质(子)结构检测。最初,它将查询结构分解为其二级结构元素(SSEs),并通过要求 SSEs 的方向和顺序具有(可调)相似性来重新组装与目标的匹配。蛋白质结构的分层组织和明智利用信息,使 deconSTRUCT 能够以提高几个数量级的速度实现已建立的搜索引擎的灵敏度和特异性,而不会以哈希形式不可挽回地束缚子结构信息。在后续处理步骤中,构建了在骨架原子水平上的匹配。呈现给用户的结果包括匹配的 SSE 列表、结构刚性叠加的变换矩阵以及几种可视化方式,这些都可以下载,也可以实现为网络浏览器插件。该服务器可在 http://epsf.bmad.bii.a-star.edu.sg/struct_server.html 上获得。