Department of Chemistry & Supercomputing Facility for Bioinformatics & Computational Biology, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India.
Phys Chem Chem Phys. 2013 Jun 21;15(23):9107-16. doi: 10.1039/c3cp44697b. Epub 2013 May 3.
We report here a novel computationally fast protocol (RASPD) for identifying good candidates for any target protein from any molecule/million molecule database. A QSAR-type equation sets up the extent of complementarity of the physico-chemical properties of the target protein and the candidate molecule and an estimate of the binding energy is generated. A correlation coefficient of 0.84 and an average error ±1.45 kcal mol(-1) are obtained for the calculated protein-ligand binding energies against experiment for more than 380 protein-ligand complexes. RASPD is seen to perform better than other popular scoring functions in predicting binding energies. The most interesting feature of this methodology is that it takes only a fraction of a second for calculating the binding energy of any ligand without docking in the active site of the target protein as opposed to several minutes for regular docking and scoring methods, while the accuracy in sorting good candidates remains comparable to that of conventional techniques. An entire million compound library, a (10(5) compound) natural product library and a (10(5) compound) NCI database can be scanned against a specified target protein within a few minutes for identifying hit molecules. The RASPD methodology is freely accessible at .
我们在此报告一种新颖的计算快速协议（RASPD），可用于从任何分子/百万分子数据库中为任何目标蛋白识别良好的候选物。一个定量构效关系型方程建立了目标蛋白和候选分子的物理化学性质互补程度，并生成了结合能的估计值。对于超过 380 个蛋白-配体复合物，计算的蛋白-配体结合能与实验值的相关系数为 0.84，平均误差为±1.45 kcal mol(-1)。RASPD 在预测结合能方面优于其他流行的评分函数。这种方法最有趣的特点是，与常规对接和评分方法需要几分钟相比，它只需几分之一秒即可计算任何配体的结合能，而无需在目标蛋白的活性位点进行对接，同时，良好候选物的排序准确性仍可与传统技术相媲美。在几分钟内，可以针对指定的目标蛋白扫描整个百万化合物库（10(5) 化合物）天然产物库和 NCI 数据库（10(5) 化合物），以识别命中分子。RASPD 方法可在 上免费获取。
