Abstract <p>A quantitative analysis of the structure–activity relationship for cyclooxygenase-2 (COX-2) inhibitors, which are diarylheterocyclic derivatives, was performed using the GUSAR 2019 program. In this work, QSAR models were constructed using the pIC<sub>50</sub> parameter [log(1/IC<sub>50</sub>)], where IC<sub>50</sub> (expressed in mmol/L) is the concentration of the drug that inhibits 50% of the enzyme activity. The experimental values of the IC<sub>50</sub> parameter for the simulated compounds ranged from 1.00–9549.93 nmol/L. Nine statistically significant consensus models were developed for predicting the numerical IC<sub>50</sub> values. These models demonstrate satisfactory prediction accuracy for the pIC<sub>50</sub> parameter for both the structures used in the training data and for two independent test sets and CHEMBL129021, CHEMBL339798, CHEMBL285831 suggests the potential use of models M1–M9 for the virtual screening of virtual libraries and databases to find new potentially efficient inhibitors of COX-2. A combination of MNA (Multilevel Neighborhoods of Atoms) and QNA (Quantitative Neighborhoods of Atoms) descriptors with three whole molecular descriptors (topological length, topological volume and lipophilicity) was used to develop 9 statistically significant, valid consensus QSAR models.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Quantitative Analysis of the Structure–Activity Relationship for Cyclooxygenase-2 Inhibitors Based on Diarylheterocyclic Derivatives

  • Yu. Z. Martynova,
  • V. R. Khairullina,
  • L. S. Maksimov

摘要

Abstract

A quantitative analysis of the structure–activity relationship for cyclooxygenase-2 (COX-2) inhibitors, which are diarylheterocyclic derivatives, was performed using the GUSAR 2019 program. In this work, QSAR models were constructed using the pIC50 parameter [log(1/IC50)], where IC50 (expressed in mmol/L) is the concentration of the drug that inhibits 50% of the enzyme activity. The experimental values of the IC50 parameter for the simulated compounds ranged from 1.00–9549.93 nmol/L. Nine statistically significant consensus models were developed for predicting the numerical IC50 values. These models demonstrate satisfactory prediction accuracy for the pIC50 parameter for both the structures used in the training data and for two independent test sets and CHEMBL129021, CHEMBL339798, CHEMBL285831 suggests the potential use of models M1–M9 for the virtual screening of virtual libraries and databases to find new potentially efficient inhibitors of COX-2. A combination of MNA (Multilevel Neighborhoods of Atoms) and QNA (Quantitative Neighborhoods of Atoms) descriptors with three whole molecular descriptors (topological length, topological volume and lipophilicity) was used to develop 9 statistically significant, valid consensus QSAR models.