Large-scale classification of P-glycoprotein inhibitors using SMILES-based descriptors

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Cancer Research

Background:

• P-glycoprotein (Pgp) plays a crucial role in multidrug resistance (MDR) in cancers.
• Inhibiting Pgp is a key strategy to overcome MDR cancers.
• Classifying Pgp inhibitors is challenging due to Pgp's substrate promiscuity.

Purpose of the Study:

• To develop a robust classification model for P-glycoprotein (Pgp) inhibitors.
• To utilize Simplified Molecular Input Line Entry System (SMILES) notations for model construction.
• To identify key chemical features influencing Pgp inhibitory activity.

Main Methods:

• Employed CORrelation And Logic (CORAL) software for classification model development.
• Utilized a dataset of 2254 Pgp inhibitors represented by SMILES notations.
• Evaluated model performance using accuracy, sensitivity, specificity, and Matthews Correlation Coefficient (MCC).

Main Results:

• The SMILES-based CORAL model demonstrated predictive performance with accuracy, sensitivity, and specificity >70%.
• The model achieved an MCC value >0.6 across training, calibration, and validation sets.
• CORAL identified specific chemical features associated with enhanced or reduced Pgp inhibition.

Conclusions:

• The CORAL software shows potential for rapid screening of potential Pgp inhibitors from large chemical libraries.
• The identified chemical features can guide the rational design of novel Pgp inhibitors.
• This approach aids in developing new strategies to combat multidrug-resistant cancers.