The application of multiple linear regression and artificial neural networks to study the quantitative structure-activity relationship of a group of chemokine derivatives

Document Type : Original Article

Authors

• mehdi nekoei ¹
• M K ²
• M M ²
• B M ³
• T B ⁴

¹ Islamic Azad University of shahrood branch

² IAU of Shahrood

³ Sabzevar University

⁴ University

Abstract

A quantitative structure-activity relationship (QSAR) study was conducted to predict the pharmacological activity of some chemokine derivatives using multiple linear regression and artificial neural networks (ANN). At first, the structure of pharmaceutical compounds was drawn and optimized with the help of Hypercam software. Then, a wide range of molecular descriptors were calculated by Dragon software. After reducing the number of descriptors that had a correlation above 0.9 and the descriptors that were more than 90% similar, stepwise regression was used to obtain the best descriptors that were most related to the pharmacological activity of the target compounds. became 7 descriptors including MATS2p, PCWTe, RDF045m, RDF065m, RDF115m, C-003 and C-040 were selected. Then, multiple linear regression (MLR) and artificial neural networks (ANN) methods were used to model and predict the activity of test series compounds. The obtained results show that both methods provide acceptable results that can be used to predict new pharmaceutical compounds.

Keywords

• Quantitative structure-activity relationship
• chemokine derivatives
• multiple linear regression
• artificial neural networks

---