Using matched molecular series as a predictive tool to optimize biological activity

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Matched molecular pairs (MMPs) are widely used to analyze structure-activity relationships (SARs).
• Matched molecular series (MMS) extend MMPs, offering a broader scope for SAR analysis by considering multiple R-group modifications at a single position.
• Existing methods for R-group prediction often lack the ability to integrate diverse SAR trends.

Purpose of the Study:

• To introduce Matsy, a novel knowledge-based method for predicting R groups that are likely to improve molecular activity.
• To evaluate Matsy's performance by comparing its predictions against the Topliss tree using ChEMBLdb activity data.
• To investigate the factors influencing the predictive success of Matsy, particularly the role of matched series length.

Main Methods:

• Development of Matsy, a knowledge-based algorithm utilizing matched molecular series to predict R-group modifications.
• Comparison of Matsy predictions with Topliss tree recommendations using curated activity data from ChEMBLdb.
• Large-scale retrospective testing to quantitatively assess the predictive performance of Matsy.

Main Results:

• Matsy demonstrates effective prediction of R groups that enhance activity based on observed activity orders within matched series.
• The predictive success of Matsy is significantly correlated with the strength of preferred orders found in matched series, especially in longer series.
• Matsy's performance was validated through extensive retrospective testing, showing its utility in drug design.

Conclusions:

• Matsy provides a powerful tool for medicinal chemists to leverage SAR trends from various projects.
• The algorithm can be applied as a Topliss-like recommendation system or a hypothesis-generation tool to guide compound design.
• Integrating information from matched molecular series enhances the prediction of R groups for improved drug activity.