External validation and prediction employing the predictive squared correlation coefficient test set activity mean vs...

Area of Science:

• * Computational chemistry and cheminformatics.
• * Drug discovery and development.

Background:

• * Quantitative structure-activity relationship (QSAR) models are crucial for predicting molecular activity.
• * External prediction capability is commonly assessed using the predictive squared correlation coefficient, q (2).
• * Current OECD guidelines recommend using the training set mean for q (2) calculation.

Purpose of the Study:

• * To investigate the accuracy of the current OECD guideline for calculating q (2) in QSAR external validation.
• * To demonstrate the systematic overestimation of predictive capability under existing guidelines.
• * To propose a revised method for calculating q (2) using the test set mean.

Main Methods:

• * Mathematical proof to demonstrate the overestimation bias.
• * Example calculations using three regression models and literature data sets.
• * Comparison of q (2) values calculated with training set mean versus test set mean.

Main Results:

• * The current OECD guideline for q (2) calculation systematically overestimates external prediction capability.
• * This overestimation is caused by differences between training and test set activity means.
• * Calculated q (2) values can exceed r (2) values for external test sets under the current method.

Conclusions:

• * The test set activity mean should be used for calculating q (2) to accurately quantify external prediction capability.
• * The OECD guidance document on QSAR external validation requires revision.
• * Accurate model evaluation is critical for reliable drug discovery and development.