|
2D-QSAR and docking studies of 4-anilinoquinazoline derivatives as epidermal growth factor receptor tyrosine kinase inhibitors
|
Mahtab Ghasemi Dogaheh    , Hesmatollah Ebrahimi-Najafabadi , Fatemeh Yousefbeyk , Saeed Ghasemi |
|
|
|
Abstract: (1152 Views) |
| Introduction: Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor derivatives play an important role in the treatment of cancer. We aim to construct 2D-QSAR models using various chemometrics using 4-anilinoquinazoline-containing EGFR TKIs. In addition, the binding profile of these compounds was evaluated using a docking study.
Materials and Methods: In this study, 122 compounds of seven different structural categories with a 4-anilinoquinazoline scaffold were obtained from the pieces of literature. 2D-QSAR models were prepared using the linear method, including multiple linear regression (MLR), alongside non-linear methods, including artificial neural networks (ANN) and support vector machines (SVM). The validation of suggested 2D-QSAR models was performed using internal and external validation techniques. For molecular docking, three compounds, including compounds 32, 75, and 98, which had the highest pIC50, were used. Molecular docking was performed using AutoDock 4.2.
Results: Selected descriptors indicated that atomic electronegativity, hydrogen bonding ability, lipophilicity, and molecular shape and volume were the factors affecting the activity of the compounds. Statistical criteria, related to interpretation and validation of the models, were in the appropriate range. The model obtained by the ANN method, with the lowest mean absolute error of 0.365 for both training and testing sets, was the best. The compound 75 as the most potent compound indicated binding energies of -8.33 kcal/mol.
Conclusion: Finally, it was found that the models could effectively predict the activity of the compounds. It was also discovered that the compounds were properly bound to the active site of the receptor. In addition, validation results showed that all processes were sufficiently valid. |
|
| Keywords: Quantitative Structure-Activity Relationship, ErbB Receptors, Regression Analysis, Support Vector Machine, Neural Networks |
|
|
Full-Text [PDF 1491 kb]
(460 Downloads)
|
Type of Study: Research |
Subject:
General
Received: 2021/06/19 | Accepted: 2021/09/19 | Published: 2022/05/20
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
