Development and OECD-Guided Experimental Validation of Eigenvalue-Based QSAR Models for Skin Permeation

Area of Science:

• Computational chemistry
• Drug discovery
• Pharmacokinetics

Background:

• Quantitative structure-activity relationships (QSARs) are crucial for drug development, but most methods rely on 2D descriptors, neglecting 3D structural information.
• Existing QSAR models often fail to capture complex interactions influencing drug properties like transdermal permeation.

Purpose of the Study:

• To develop and validate a novel QSAR methodology, EigenValue Analysis (EVANS), integrating 2D and 3D molecular descriptors for enhanced prediction of transdermal permeation.
• To build and experimentally validate a predictive model for human skin permeability using the EVANS approach.

Main Methods:

• Developed the EVANS methodology, a hybrid QSAR approach combining 2D physicochemical properties and 3D geometric descriptors.
• Curated a high-quality dataset of 187 compounds with human skin permeability data.
• Built and optimized QSAR models using MLR, RF, and SVM algorithms, with SVM demonstrating superior performance.

Main Results:

• The SVM-based EVANS model, utilizing five hybrid descriptors, achieved the best performance in predicting logarithmic skin permeability coefficient (log Kp).
• External validation using the SkinPiX database and experimental testing of two compounds (4-methylumbelliferone, levofloxacin) confirmed the model's generalisability and accuracy.
• The EVANS model predictions closely matched experimental values and outperformed established transdermal permeation models.

Conclusions:

• The EVANS methodology provides a robust and experimentally validated approach for predicting transdermal permeation.
• This novel QSAR approach enhances drug-likeness prescreening by incorporating crucial 3D structural information.
• The open-source EVANS methodology offers a valuable tool for drug discovery and development, improving the accuracy of pharmacokinetic predictions.