標題: DS-SARST: 蛋白質結構區域交換現象之高效能搜尋比對引擎
DS-SARST: an efficient protein structure database searching server for three-dimensional domain swapping proteins
作者: 曾靜怡
Tseng,Ching-Yi
羅惟正
Lo, Wei-Cheng
生物資訊及系統生物研究所
關鍵字: 蛋白質結構區域交換;蛋白質結構比對;蛋白質四級結構;3D domain swapping;protein structure comparison;protein quaternary structure
公開日期: 2013
摘要: 蛋白質結構區域交換現象之高效能搜尋比對引擎,名為DS-SARST (Three-dimensional Domain Swapping Search Aided by Ramachandran Sequential Transformation), DS-SARST是一個可以在幾分鐘內快速掃描蛋白質資料庫 (Protein Data Bank) 並收集相關具有蛋白質結構區域交換現象的同源蛋白質的搜尋引擎。現今的結構比對方法都有不同的缺點存在,準確率快的方法需要花時間,反之,能快速比對的方法,準確度卻很低。如果將兩種方法結合,以互相彌補缺點,也許能夠變成一個兼具效率跟準確的新方法,DS-SARST便是如此產生。我們採取“filter-and refine”策略,在“screening stage”,先利用RST (Ramachandran Sequential Transformation) 線性演算法,將蛋白質從三維結構轉換成一維線性字串現象的演算法ADiDoS (Angle-Distance image-based determination of three-dimensional Domain Swapping proteins) 進而篩選出確實符合結果的蛋白質。雖然SARST在辨認具有蛋白質結構區域交換現象的蛋白質能力比ADiDoS差,可是在比對速度上卻比ADiDoS快上好幾千倍。在“refinement stage”,利用專門比對蛋白質結構區域交換現象的演算法,ADiDoS,來降低screening stage篩選出的蛋白錯誤率。 藉由結合這兩種方法並互相彌補不足,DS-SARST成為一個能夠快速且準確偵測有此現象的蛋白質搜尋引擎。整體來說,DS-SARST計算過程不用三分鐘,即便使用者關閉瀏覽器,之後依舊可以使用session ID 取回計算結果。最後結果頁會列出所有可能有蛋白質結構區域交換現象的蛋白和其相關資料供使用者參考。 簡言之,DS-SARST是一個創新且專門用來偵測蛋白質結構區域交換現象的高效能搜尋引擎。我們相信DS-SARST對於未來研究蛋白質結構區域交換現象必定是個有用的工具,且對於日後發展相關資料庫一定也會有很大的幫助。  
DS-SARST (three-dimensional Domain Swapping Search Aided by Ramachandran Sequential Transformation) is a web service for identifying three-dimensional domain swapping (hereafter abbreviated as DS) proteins from large structure databases. It works based on a “filter-and-refine” strategy. The filtering engine of DS-SARST is SARST (Structural similarity search Aided by Ramachandran Sequential Transformation), a rapid protein structural similarity search method more sensitive than conventional structure comparison methods in detecting DS relations between proteins; the refinement engine is ADiDoS (Angle-Distance image-based determination of three-dimensional Domain Swapping proteins), the state-of-the-art DS-detection algorithm. Although the specificity of SARST in identifying DS-related homologs is much lower than that of ADiDoS, it runs a million times faster than ADiDoS. By combining these two methods, DS-SARST can precisely and rapidly detect DS cases. The calculation takes less than 3 minutes, DS-SARST achieves a precision almost the same as ADiDoS and a scanning speed hundreds of thousands times higher. To sum up, DS-SARST is an innovative web service for detecting DS proteins. We expect that DS-SARST can be helpful for the understanding DS and will facilitate the application and research development of three-dimensional domain swapping in biotechnology fields.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070157206
http://hdl.handle.net/11536/74387
顯示於類別:畢業論文