基于预测模型优化的药物分子对接网格

Drug molecular docking design grid based on optimized forecast models

将网格技术应用于药物分子对接中,能有效解决药物分子对接中所涉及的搜索空间巨大、时间耗费长、对计算环境要求高等问题.通过应用改进的遗传算法多种群竞争机制的对接演化模型GAsDock,以信息熵控制设计空间的收缩,增强了进化的目的性,显著地提高了对接效率.然后,利用线性、指数平滑和离散马尔可夫3种预测模型,并结合加权系数法,提高了分子对接任务的效率及各节点计算能力的自适应性.同时,采用了网格数据传输优化技术,降低了药物分子的传输时间,且能更有效地利用网格资源.实例测试表明了药物分子对接与网格技术相结合的合理性及有效性.

Rigid conditions,such as searching space,spending time and computing environment,are required by drug molecular docking design,which can be satisfied through combining drug molecular docking design with the techniques of grid.An improved genetic algorithm docking model,GAsDock,is presented,which is based on multi-population competition mechanism.And the information entropy is used to control the solution space narrowing down,so the evolutional reasonability and docking efficiency are all improved greatly.Then three forecast models,such as linear method,exponential flatness method and discrete Markov method,are adopted.Combined with weighted coefficients method,the efficiency of molecule docking tasks and the self-adapted capabilities of nodes are all improved.Simultaneously,optimized technique of grid data-transferring is adopted,through which,the docking time is reduced and the grid resources are used more efficiently.The results of example test show that the combination of drug molecular docking design and grid techniques is efficient and reasonable.