作者: Sean Ekins , Shelly Binkley , Barbara J. Ring , Steven A Wrighton , Jennifer S. Gillespie
DOI:
关键词:
摘要: The program Catalyst was used to build three-dimensional quantitative structure activity relationship (3D-QSAR) pharmacophore models of the structural features common competitive-type inhibitors cytochrome P-450 (CYP) 3A4. These were compared with 3D- and four-dimensional (4D)-QSAR partial least-squares (PLS) built using molecular surface-weighted holistic invariant (MS-WHIM) descriptors for size shape inhibitor. model generated from multiple conformers competitive CYP3A4-mediated midazolam 1'-hydroxylation (n = 14) yielded a high correlation observed predicted Ki values r 0.91. Similarly, PLS MS-WHIM produce 4D-QSARs this data set produced that statistically predictable after cross-validation. Two additional pharmacophores constructed literature 32) derived inhibition CYP3A-mediated cyclosporin A metabolism IC50 22) quinine 3-hydroxylation. illustrated correlations CYP3A4 0.77 0.92, respectively. corresponding by these sets comparable quality as judged Both also validated predicting Ki(apparent) test eight not included in either model. In seven cases, residuals within 1 log unit values. 4D-QSAR study suggest utility future silico prediction drug-drug interactions.