Causal-Chemprop: Causal Machine Learning for Molecular Property Prediction and Design

Abstract

A priori estimation of molecular properties has long been a topic of immense interest to the pharmaceutical sciences for molecular design. While neural network-based models have achieved high predictive accuracy, they have limited utility in molecular design, particularly in small and out-of-distribution data regimes. Current neural network-based models lack mechanisms to explicitly incorporate prior knowledge from existing data. Herein, we introduce a causal machine learning framework built on the Chemprop and DAGMA architectures for molecular property prediction called Causal-Chemprop. Its capacity for counterfactual reasoning supports human-in-the-loop optimization of molecular structure, which we demonstrate by improving the rank-ordering of out-of-distribution kinase inhibitor pIC₅₀ values─most where standard encoders extrapolate poorly─both within and across scaffold series, in a lead-optimization setting in which the test scaffold is known. We further integrate Causal-Chemprop with the molecular optimization algorithm EvoMol as a scoring function, where it tends to yield more synthesizable molecules that more closely resemble the sought-after active compound than Chemprop. To our knowledge, this is the first application bridging neural network-learned molecular representations with structural causal models for molecular property prediction and optimization via counterfactual reasoning, suggesting several promising new avenues for exploration in molecular property prediction and design.