Predictive modeling of asthma drug properties using machine learning and topological indices in a MATLAB based QSPR...

Area of Science:

• Computational chemistry and cheminformatics
• Pharmacology and drug discovery

Background:

• Machine learning (ML) is crucial for predicting compound properties in drug development.
• Quantitative Structure-Property Relationship (QSPR) studies integrate molecular structure with biological activity.
• Asthma drug discovery requires accurate prediction of physio-chemical properties.

Purpose of the Study:

• To evaluate the predictive power of topological indices combined with ML algorithms for asthma drug physio-chemical properties.
• To explore the utility of random forest and extreme gradient boosting in this context.
• To highlight the potential of computational strategies in pharmaceutical research.

Main Methods:

• Utilized MATLAB-based algorithms to compute topological indices.
• Applied machine learning algorithms, including random forest and extreme gradient boosting.
• Trained and validated models using labeled data for property prediction.

Main Results:

• Demonstrated the ability of ML algorithms to accurately predict physio-chemical properties of asthma drugs.
• Showcased the synergy between topological indices and ML for precise drug structure analysis.
• Validated the efficiency of ML in evaluating large datasets.

Conclusions:

• The integration of ML with topological indices significantly enhances the accuracy of predicting drug properties.
• Computational strategies, particularly ML, offer a powerful and efficient approach to pharmaceutical discovery.
• This study facilitates the development of novel and improved asthma medications.