標題: | 利用ICP演算法之蛋白質結構比對工具 Protein Structure Alignment Tool by ICP Algorithm |
作者: | 羅珮芬 Pei-Fen Lo 荊宇泰 Yu-Tai Ching 資訊科學與工程研究所 |
關鍵字: | 結構比對;蛋白質;structure alignment;protein |
公開日期: | 2002 |
摘要: | 在生物資訊的領域中,經常會利用比對兩個基因序列或蛋白質以分析兩者之間的差異。隨著後基因體時代的來臨,生物學家相信蛋白質的功能與其結構有相當的關聯性。蛋白質結構比對便成為生物學家以結構的角度來觀察各蛋白質間相關程度的工具。本論文利用Besl與McKay在1992年提出之ICP演算法,應用於蛋白質結構上,藉由尋找兩組蛋白質間距離最近的對應關係,計算其幾何轉換矩陣,包括旋轉矩陣與轉置矩陣,反覆縮小兩組蛋白質之RMSD(Root Mean Square Deviation),最後將收斂得到最佳的幾何轉換矩陣,將兩組蛋白質對齊在一起。另外,發展擷取蛋白質的整體特徵以及拓樸順序的限制等機制來加快計算速度及搜尋理想的比對結果。利用所計算出的幾何轉換矩陣繪出兩組蛋白質的疊合狀況,可自由以滑鼠旋轉觀察蛋白質之間實際的比對情形。
本論文所採用的實驗資料來源是由PDB(Protein Data Bank)所擷取下來。論文中實驗的蛋白質數據大部分來自於所參考的文獻。由實驗得知,本論文所提出的方法與機制可以正確分析蛋白質的相似結構與程度。同時提供方便使用的介面及蛋白質視覺化的功能。 In bioinformatics, it is desired to compare differences by aligning two proteins structure or nucleotide sequences. Biologists then become aware of the relationship from structures to functions for proteins. The protein structure alignment provides biologists observing any two proteins’ relationships. In this thesis, we used the ICP algorithm proposed by Besl and McKay at 1992 on protein structure analysis. ICP algorithm iteratively minimizes the distance between two different protein structures and computes a transformation matrix. Finally an optimal transformation matrix is obtained to align two proteins. In this thesis, we also present methods to speed up computing time. The experiment data of this thesis are obtained from PDB (Protein Data Bank). As a result of experiments, we can analysis protein structures accurately and provide friendly interface for protein visualization. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#NT910394096 http://hdl.handle.net/11536/70262 |
Appears in Collections: | Thesis |