基于蚁群算法的计算机辅助药物设计研究

发布时间：2018-01-24 16:38

本文关键词： 分子对接蚁群算法函数优化计算机辅助药物设计　出处：《电子科技大学》2016年硕士论文　论文类型：学位论文

【摘要】：成本投入高、开发风险大、研制周期长一直都是药物设计和研发所面临的难题,这些困难对新药研制的发展有着很大的影响。随着当今计算机技术的迅猛发展,计算机辅助药物分子的研究模型也越来越成熟。在当今药物分子的探索和研制中,一种基于结构化的药物分子模拟方法-分子对接技术,已经逐渐成为创新药物研制的重要领域之一,同时也是先导化合物的探索方向的重要途径。分子对接技术是根据两个已知三维立体结构的单体生物分子(在药物设计中,通常是大分子蛋白质和小分子配体),通过计算机来对这两个单体分子复合物的结合模型进行仿真。分子对接技术作为计算机辅助药物分子设计的重要组成部分,不仅具有计算机的高速计算能力,还能有效地结合优化方法并使之与药物研制相融合。蚁群算法(Ant Colony Optimization,ACO)作为一种模拟自然界中蚂蚁觅食行为的新型群智能优化算法,具有较强的可扩充性、系统健壮性好、能进行并行分布式计算、易于同其他仿生算法相结合等优点。因此,本文将对蚁群算法进行优化并将这种改进的算法应用于分子对接中,并将蚁群算法与构想搜索中的全局搜索过程中所常用的遗传算法进行性能上的比较,可以预见将蚁群算法与常用的分子对接平台AutoDock整合到一起是非常具有实际意义的。将蚁群算法收敛速度快,对接精度高的特点与分子对接软件平台AutoDock结合起来,对分子对接过程中的全局搜索进行替代,从而最终达到对分子对接过程进行优化的目的。本文的主要工作是保留使用AutoDock4.2平台中的打分函数和柔性策略,将改进的蚁群算法作为对接过程中的构象搜索策略,代替软件中原有的遗传算法,通过对蛋白质和配体的对接仿真验证其优越性。分子对接过程,实际上就是对一个复杂的函数模型进行优化的过程,而改进的蚁群算法则具有很强的自适应特点,能应用于大部分优化问题中,因此在药物设计中也表现不错。本文旨在将改进的蚁群算法同常用对接软件AutoDock相结合,通过具体蛋白质分子同配体分子之间的结合测试来对其进行仿真实验并在此基础上进行分析,通过与典型的遗传算法进行试验对比突出蚁群算法在一些方面的优越性,并在此基础上对具有国际影响力的高性能分子对接软件进行完善。最终可以得出结论:基于蚁群算法的分子对接平台在对对接程中产生的庞大构象进行搜索时具有更快的收敛速度和更好的效率。
[Abstract]:With the rapid development of computer technology, high cost, high risk and long research cycle have been the problems of drug design and development. These difficulties have a great impact on the development of new drug development. The research model of computer-aided drug molecules is becoming more and more mature. In the exploration and development of drug molecules, a molecular docking method based on structured drug molecular simulation is proposed. It has gradually become one of the important fields in the development of innovative drugs. Molecular docking technology is based on two known three-dimensional structures of monomer biomolecules (in drug design). Usually macromolecular proteins and small molecular ligands. As an important part of computer aided drug molecular design, molecular docking technology not only has the high speed computing ability of computer, but also simulates the binding model of these two monomers by computer. Ant Colony Optimization can be effectively combined with optimization method and combined with drug development. ACO) is a new swarm intelligence optimization algorithm which simulates the foraging behavior of ants in nature. It has strong scalability, good system robustness, and can be used for parallel and distributed computing. It is easy to combine with other bionic algorithms. Therefore, this paper optimizes the ant colony algorithm and applies the improved algorithm to molecular docking. The performance of ant colony algorithm is compared with that of genetic algorithm which is commonly used in global search. It can be predicted that the integration of ant colony algorithm with the commonly used molecular docking platform AutoDock is of great practical significance, and the convergence rate of ant colony algorithm is fast. The characteristics of high docking precision are combined with the molecular docking software platform AutoDock to replace the global search in the molecular docking process. In order to optimize the docking process, the main work of this paper is to retain the scoring function and flexible strategy in the AutoDock4.2 platform. The improved ant colony algorithm is used as the conformational search strategy in the docking process instead of the original genetic algorithm in the software. The simulation of protein and ligand docking proves its superiority and molecular docking process. In fact, it is the process of optimizing a complex function model, and the improved ant colony algorithm has strong adaptive characteristics, and can be applied to most optimization problems. This paper aims to combine the improved ant colony algorithm with the commonly used docking software AutoDock. Through the binding tests between specific protein molecules and ligand molecules, the simulation experiments were carried out and the analysis was carried out on this basis. Compared with the typical genetic algorithm, this paper compares the advantages of ant colony algorithm in some aspects. On this basis, the high performance molecular docking software with international influence is perfected. Finally, the conclusion can be drawn:. The molecular docking platform based on ant colony algorithm has faster convergence speed and better efficiency in searching the huge conformation generated in docking process.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18;R91

【相似文献】