Development of QSRR model for hydroxamic acids using PCA-GA-BP algorithm incorporated with molecular...

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Analytical Chemistry

Background:

• Hydroxamic acids are potent zinc chelators used in designing inhibitors for zinc metalloenzymes like histone deacetylases (HDACs).
• Quantitative Structure-Retention Relationships (QSRR) studies are crucial for predicting chromatographic behavior based on molecular properties.

Purpose of the Study:

• To develop a QSRR model for predicting the retention time of hydroxamic acids in High-Performance Liquid Chromatography (HPLC).
• To evaluate the utility of molecular interaction-based features, specifically molecular docking scores, as novel descriptors in QSRR modeling.

Main Methods:

• A series of hydroxamic acids with HDAC inhibitory activities were analyzed using QSRR.
• Principal Component Analysis (PCA) was employed for descriptor dimension reduction.
• A Genetic Algorithm-based Error Backpropagation (GA-BP) algorithm was used with principal components.
• Double cross-validation was performed for model validation.
• Molecular docking scores were introduced as molecular interaction-based features.

Main Results:

• The developed QSRR model demonstrated good predictive performance.
• Incorporation of molecular interaction-based features significantly enhanced the model's accuracy.
• The final model achieved an R-squared value of 0.842, RMSEP of 0.440, and MAE of 0.573.

Conclusions:

• A robust QSRR model was successfully developed for predicting hydroxamic acid retention times in HPLC.
• Molecular interaction-based features, such as docking scores, are feasible and effective descriptors for improving QSRR model accuracy.
• This study provides a valuable tool for chromatographic analysis and contributes to understanding structure-retention relationships.