David L, Luo R, Gilson M K
Center for Advanced Research in Biotechnology, Rockville, MD 20850, USA.
J Comput Aided Mol Des. 2001 Feb;15(2):157-71. doi: 10.1023/a:1008128723048.
The optimizer developed for the Mining Minima algorithm, which uses ideas from Genetic Algorithms, the Global Underestimator Method, and Poling, has been adapted for use in ligand-receptor docking. The present study describes the resulting methodology and evaluates its accuracy and speed for 27 test systems. The performance of the new docking algorithm appears to be competitive with that of previously published methods. The energy model, an empirical force field with a distance-dependent dielectric treatment of solvation, is adequate for a number of test cases, although incorrect low-energy conformations begin to compete with the correct conformation for larger sampling volumes and for highly solvent-exposed binding sites that impose little steric constraint on the ligand.
为采矿极小值算法开发的优化器采用了遗传算法、全局低估方法和极化的思想，已被改编用于配体-受体对接。本研究描述了由此产生的方法，并评估了其在27个测试系统中的准确性和速度。新对接算法的性能似乎与先前发表的方法具有竞争力。能量模型是一种对溶剂化采用距离依赖性介电处理的经验力场，对于许多测试案例来说是足够的，尽管对于更大的采样体积和对配体几乎没有空间约束的高度溶剂暴露的结合位点，不正确的低能量构象开始与正确构象竞争。
