結構生物學分析系統：以GA/GP增進藥物對接（嵌合）模擬準確率

Analysis System of Structural Biology： Improved Accuracy Rate of Drug Docking Simulation with GA/GP (Genetic Algorithms/ Genetic Programming)

王啟華(Chi-Hua Wang)

指導教授 ： 龐金宗

共同指導教授 ： 樓國隆(Kuo-Long Lou)

元智大學/資訊學院/資訊管理學系/碩士(2006年)

https://doi.org/10.6838/YZU.2006.00112

摘要

傳統的製藥過程，是一件相當浪費時間與成本的工作。一般來說，都必須經過四個階段：初始分子結構、精製分子結構以產生新藥、生化測試（biological test）、臨床測試（clinical test）。每一個階段大都需要三年到五年的時間。
近年來由於資訊科技的進步，利用電腦計算的能力，以期望減少前置階段所花費的時間。電腦輔助藥物設計（CADD，Computer Aided Drug Design）就是主要利用的方法之一。利用電腦技術去模擬蛋白質與藥物分子的結合點，進一步的去篩選兩者結合的可能性。不但能有效節省藥物開發過程中，所花費的時間與成本，並且還能更進一步的利用分子結構，去瞭解蛋白質本身構成時所隱含的各種關係。
但是在這個尋找的過程，所必須考慮的因素相當的多，相對就必須要有一個快速且有力的演算法或是技術去進行這項繁複的工作。基因演算法（GA，Genetic Algorithms）與基因規劃（GP，Genetic Programming）提供了一個解決這個問題的途徑。然而不論是基因演算法或是基因規劃都需要一個好的評分方程式來進行世代評分的工作。XSCORE提供了一個、高準確率、高彈性的選擇。
本研究針對蛋白質與藥物分子結合點尋找的問題。利用基因演算法與基因規劃進行結合點尋找的工作，並且使用XSCORE擔任評分方程式的任務。根據本研究的研究結果發現，基因演算法與基因規劃的確可以有效的找到蛋白質與藥物分子的結合點。本研究的研究結果也提供了未來在藥物開發過程中，一個可供參考的依據。

關鍵字

結構生物學 ； 基因演算法 ； 基因規劃 ； 分子對接 ； XSCORE

並列摘要

Conventional procedures for drug design have been very expensive and time-consuming. In general, there are four phases regarding such procedures: initiation of molecular structure, optimization of the structure, biochemical investigations, and clinical trials. Each phase may take approximately three to five years to produce a new drug.
Due to the tremendous progresses on information technology during recent years, it is expected to shorten the required research time spent in the early period of the aforementioned development through computer calculation. CADD (Computer Aided Drug Design) is one of the most powerful concepts applied to satisfy such demand. Upon docking simulations, it is allowed to find out the binding sites and orientations between target proteins and drug molecules in several days. This is not only to save the time and the cost used in drug development, but also for us able to understand the structural implications used for further design. However, it is still currently difficult to formularize efficient software to carry out the docking simulations as a standard procedure leading to definite results with high accuracy. Therefore, this study is in attempts to propose a new category of programming, for which the standard effectiveness for docking procedure can be anticipated in the near future.
To initiate such computer simulations, many factors have to be taken into consideration. The first is to decide which algorithms should be applied to perform the job. GA (Genetic Algorithms) and GP (Genetic Programming) seem to be excellent candidates to solve this problem. The next concern is the determination of scoring function, which is appropriate for either GA or GP to generate their scores. As being the best commercially available scoring function with high accuracy and flexibility, XSCORE is used to satisfy this purpose.
In addition, we concentrate our study in the search of binding site(s) between protein and the drug molecule through docking simulations by applying the aforementioned special algorithms and scoring function. At present stage, target protein is regarded as a rigid body, whereas the drug molecule is allowed to be entirely flexible. According to our results, GA and GP can indeed achieve the searching of correct docking sites between target protein and the drug molecules. This finding seems to be critical in giving evidence of the application of our new method in the drug design procedure. Further studies with partially flexible protein regions added into the docking system will be anticipated in the future and thus accelerate the development of effective software for the design of new lead compounds.

並列關鍵字

docking ； genetic algorithm ； genetic programming ； XSCORE

參考文獻

[1]Bäck, Thomas ,1996 , “Evolutionary Algorithms in Theory and Practice”, Oxford

[2]Goodsell, D. S., Morris,G. M., Olson, A. J. , 1996, “Docking of flexible ligands: applications of AutoDock”. J. Med. Chem.

[3]Holland, John H. , 1992,”Adaptation in Natural and Artificial Systems”, the MIT

[5]Maria Kontoyianni, Laura M. McClellan, and Glenn S. Sokol, 2004, “Evaluation of Docking Performance: Comparative Date on Docking Algorithms”, J. Med. Chem, 47, 558-565.

[6]Morris, G. M., Goodsell, D. S., Halliday, R. S., Huey, R., Hart, W. E., Belew, R. K., andOlson, A. J. , 1998 . “Automated docking using a lamarckian genetic algorithm and empiricalbinding free energy function.” , J. Comput. Chem. 19, 1639-1662. Press, Cambridge. ecognition 9, 1-5.