Perekhodtsev Gleb D
Algodign LLC, entrance 4, Bolshaya Sadovaya street 8, Moscow 123001, Russia.
Mol Divers. 2006 Feb;10(1):81-3. doi: 10.1007/s11030-006-2153-0.
Two-dimensional structural similarity calculations have been applied to estimate binding affinities of serine proteases inhibitors. 1103 trypsin binders, 1268 thrombin binders and 714 fXa binders have been used to compare experimental and predicted data. The predictions generally provide reasonable estimates of the observed binding affinities. The accuracy of the predictions depends on the size of the dataset, but not dramatically. The accuracy also depends on the number of similar structures used to make the calculations, with a number from 3 to 6 usually being optimal. The binding affinity is noticeably more sensitive to the inhibitor structure in the case of fXa relative to thrombin.
二维结构相似性计算已被用于估算丝氨酸蛋白酶抑制剂的结合亲和力。使用了1103种胰蛋白酶结合剂、1268种凝血酶结合剂和714种因子Xa结合剂来比较实验数据和预测数据。这些预测通常能合理地估计观察到的结合亲和力。预测的准确性取决于数据集的大小,但影响不大。准确性还取决于用于进行计算的相似结构的数量,通常3到6个数量是最佳的。与凝血酶相比,因子Xa的情况下,结合亲和力对抑制剂结构的敏感性明显更高。