A Hybrid Structure-Based Machine Learning Approach for Predicting Kinase Inhibition by Small Molecules

Abstract

Kinases have been the focus of drug discovery programs for three decades leading to over 70 therapeutic kinase inhibitors and biophysical affinity measurements for over 130,000 kinase-compound pairs. Nonetheless, the precise target spectrum for many kinases remains only partly understood. In this study, we describe a computational approach to unlocking qualitative and quantitative kinome-wide binding measurements for structure-based machine learning. Our study has three components: (i) a Kinase Inhibitor Complex (KinCo) data set comprising in silico predicted kinase structures paired with experimental binding constants, (ii) a machine learning loss function that integrates qualitative and quantitative data for model training, and (iii) a structure-based machine learning model trained on KinCo. We show that our approach outperforms methods trained on crystal structures alone in predicting binary and quantitative kinase-compound interaction affinities; relative to structure-free methods, our approach also captures known kinase biochemistry and more successfully generalizes to distant kinase sequences and compound scaffolds.

Guided Docking as a Data Generation Approach Facilitates Structure-Based Machine Learning on Kinases

“…The focus on certain kinases let to large amounts of data for single kinases, such as EGFR and PKA (see Figure ). Note that the issue of uncharted kinase X-ray structure territory could be overcome by using structural modeling techniques. , In KinCo, they generated 137,778 kinase-ligand complexes using blind docking into kinase homology models. While this way, they can cover the complete kinome tree, the blind docking process might suffer in accuracy.…”

“…As of December 2023, this includes 19,443 QM and 16,972 MD simulations. Other studies investigate augmenting the structure space using AlphaFold or homology modeling, with two recent examples from the kinase field. , The KinCo dataset focuses on kinase inhibitors and provides docked poses for 137,778 kinase-ligand complexes, annotated with bioactivity measurements. The complexes are generated using blind docking into kinase homology models covering the complete kinome tree.…”

Guided docking as a data generation approach facilitates structure-based machine learning on kinases

“…Note that the issue of uncharted kinase x-ray structure territory could be overcome by using structural modelling techniques. 24,25 In KinCo, they generated 137 778 kinase-ligand complexes using blind docking into kinase homology models. While this way, the can cover the complete kinome tree, the blind docking process might suffer in accuracy.…”

“…Other studies investigate into augmenting the structure space using AlphaFold or homology modelling, with two recent examples from the kinase field. 24,25 The KinCo dataset 24 focuses on kinase inhibitors and provides docked poses for 137 778 kinase-ligand complexes, annotated with bio-activity measurements. The complexes are generated using blind docking into kinase homology models covering the complete kinome tree.…”