Boosting Sinh Cosh Optimizer and arithmetic optimization algorithm for improved prediction of biological activities...

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Quantitative Structure Activity Relation (QSAR) models are crucial for drug discovery and hazard evaluation.
• Existing QSAR models require enhancement for accurate prediction of biological activities.
• Indoloquinoline derivatives show potential as anticancer agents.

Purpose of the Study:

• To improve the predictive power of QSAR models for indoloquinoline derivatives against various cancer cell lines.
• To develop a modified machine learning (ML) technique incorporating feature selection.
• To evaluate the efficacy of the proposed model in predicting anticancer activity.

Main Methods:

• A modified machine learning technique combining the Arithmetic Optimization Algorithm (AOA) and Sinh Cosh Optimizer (SCHO).
• The AOA operators were utilized to enhance SCHO performance and act as a feature selection mechanism.
• The model was trained and validated using a dataset of indoloquinoline derivatives against human cancer cell lines (MV4-11, HCT116, A549).

Main Results:

• The proposed model achieved low root mean square error (RMSE) values: 0.6822 for MV4-11, 0.6787 for HCT116, and 0.4411/0.4477 for A549 cell lines.
• The model demonstrated high accuracy in predicting pIC50 values for indoloquinoline derivatives.
• Comparison with established methodologies indicated superior performance of the suggested model.

Conclusions:

• The developed ML-enhanced QSAR model accurately predicts the anticancer activity of indoloquinoline derivatives.
• This approach holds significant promise for identifying novel anticancer drug candidates.
• The study validates the biological application of indoloquinoline derivatives as potential therapeutics against human cancers.