Inferring molecular inhibition potency with AlphaFold predicted structures

Abstract

Even though in silico drug ligand-based methods have been successful in predicting interactions with known target proteins, they struggle with new, unassessed targets. To address this challenge, we propose an approach that integrates structural data from AlphaFold 2 predicted protein structures into machine learning models. Our method extracts 3D structural protein fingerprints and combines them with ligand structural data to train a single machine learning model. This model captures the relationship between ligand properties and the unique structural features of various target proteins, enabling predictions for never before tested molecules and protein targets. To assess our model, we used a dataset of 144 Human G-protein Coupled Receptors (GPCRs) with over 140,000 measured inhibition constants (K_i) values. Results strongly suggest that our approach performs as well as state-of-the-art ligand-based methods. In a second modeling approach that used 129 targets for training and a separate test set of 15 different protein targets, our model correctly predicted interactions for 73% of targets, with explained variances exceeding 0.50 in 22% of cases. Our findings further verified that the usage of experimentally determined protein structures produced models that were statistically indistinct from the Alphafold synthetic structures. This study presents a proteo-chemometric drug screening approach that uses a simple and scalable method for extracting protein structural information for usage in machine learning models capable of predicting protein-molecule interactions even for orphan targets.

Keywords: In silico drug discovery; Machine learning; Protein structure; Proteo-chemometrics; Quantitative structure-activity relationship modeling (QSAR); Structure based virtual screening.

Figure 1

Visual representation of the traditional QSAR approach (a) where the descriptors of the molecules with known activities are used to create a model capable of predicting activities for untested molecules. On the left (b), is a visual representation of the proposed proteo-chemometric methodology where fingerprints for multiple targets and molecules are used to create the model.

Figure 2

Generating proteins structural fingerprints from close amino acids.

Figure 3

The three modeling approaches followed. (A) Baseline individual QSAR models without any target based information; (B) Unified model with both ligand and target structure data with randomly selected data from all targets for validation; (C) Unified Blind model built with data from a subset of targets using data from an unknown target for validation.

Figure 4

Comparing validation results of Baseline QSAR models and Unified Model for Root Mean Squared Error (RMSE) and Ratio of Variance Explained (RVE) for 144 targets.