酶-配体复合物亲和性的计算

Estimating Binding Affinities for Enzyme-Ligand Complexes

描述了一种新的计算酶－配体复合物亲和性的方法．它考虑了酶－配体结合过程中自由能变化的各主要因素，并利用经验公式加以计算、从蛋白质结构数据库中选取了66个酶－配体复合物作为训练集，利用回归分析得出最后的模型．此模型通用于各种类型的酶－配体复合物，计算结果比技准确，预测算合物解离常数的平均偏差小于一个数量级．此方法还可以定量评价配体分子中每个部分对结合过程的贡献大小。

A new method is presented to estimate the binding affinity for a given enzyme-ligand complex of known three-dimensional structure This method SXORE,uses empirical scoring func-tion to describe the free energy of the binding process,which mainly accounts for enzyme-ligand interaction desolvation, and deformation effect A diverse training set of 66crystalline complexeswas analyzed by regressional statistics to obtamed the final model The model satisfactorily repro-duced the dissociation constants of the tranining set with a standard deviation of 0.86log units..A major innovation of this method is the introduction of atomic binding score This makes it a valuable tool for structure-based quantitative structure-activity relationship studies.