Rome Center for Molecular Design, Dipartimento di Chimica e Tecnologie del Farmaco, Sapienza Università di Roma, Rome, Italy.
J Comput Aided Mol Des. 2012 Aug;26(8):907-19. doi: 10.1007/s10822-012-9586-6. Epub 2012 Jul 26.
An enhanced version of COMBINE that uses both ligand-based and structure-based alignment of ligands has been used to build a comprehensive 3-D QSAR model of wild-type HIV-1 reverse transcriptase and drug-resistant mutants. The COMBINEr model focused on 7 different RT enzymes complexed with just two HIV-RT inhibitors, niverapine (NVP) and efavirenz (EFV); therefore, 14 inhibitor/enzyme complexes comprised the training set. An external test set of chiral 2-(alkyl/aryl)amino-6-benzylpyrimidin-4(3H)-ones (DABOs) was used to test predictability. The COMBINEr model MC4, although developed using only two inhibitors, predicted the experimental activities of the test set with an acceptable average absolute error of prediction (0.89 pK (i)). Most notably, the model was able to correctly predict the right eudismic ratio for two R/S pairs of DABO derivatives. The enhanced COMBINEr approach was developed using only software freely available to academics.
已经使用基于配体和基于结构的配体对齐的 COMBINE 增强版本来构建野生型 HIV-1 逆转录酶和耐药突变体的综合 3-D QSAR 模型。COMBINEr 模型专注于与仅两种 HIV-RT 抑制剂（奈韦拉平 (NVP) 和依法韦仑 (EFV））结合的 7 种不同的 RT 酶；因此，14 个抑制剂/酶复合物构成了训练集。手性 2-(烷基/芳基)氨基-6-苄基嘧啶-4(3H)-酮 (DABO) 的外部测试集用于测试可预测性。尽管仅使用两种抑制剂开发的 COMBINEr 模型 MC4 预测实验活性，但具有可接受的平均绝对预测误差（0.89 pK(i)）。值得注意的是，该模型能够正确预测 DABO 衍生物的两对 R/S 对的正确对映比。仅使用学术人员免费获得的软件开发了增强型 COMBINEr 方法。
