A large scale benchmark for molecular machine learning consisting of multiple public datasets, metrics, featurizations and learning algorithms.

Summary Activation of the μ-opioid receptor (μOR) is responsible for the efficacy of the most effective analgesics. To understand the structural basis for μOR activation, we obtained a 2.1 Å X-ray crystal structure of the μOR bound to the morphinan agonist BU72 and stabilized by a G protein-mimetic camelid-antibody fragment. The BU72-stabilized changes in the μOR binding pocket are subtle and differ from those observed for agonist-bound structures of the β2 adrenergic receptor (β2AR) and the M2 muscarinic receptor (M2R). Comparison with active β2AR reveals a common rearrangement in the packing of three conserved amino acids in the core of the μOR, and molecular dynamics simulations illustrate how the ligand-binding pocket is conformationally linked to this conserved triad. Additionally, an extensive polar network between the ligand-binding pocket and the cytoplasmic domains appears to play a similar role in signal propagation for all three GPCRs.

Chemokines are small proteins that function as immune modulators through activation of chemokine G protein–coupled receptors (GPCRs). Several viruses also encode chemokines and chemokine receptors to subvert the host immune response. How protein ligands activate GPCRs remains unknown. We report the crystal structure at 2.9 angstrom resolution of the human cytomegalovirus GPCR US28 in complex with the chemokine domain of human CX3CL1 (fractalkine). The globular body of CX3CL1 is perched on top of the US28 extracellular vestibule, whereas its amino terminus projects into the central core of US28. The transmembrane helices of US28 adopt an active-state–like conformation. Atomic-level simulations suggest that the agonist-independent activity of US28 may be due to an amino acid network evolved in the viral GPCR to destabilize the receptor’s inactive state.

The arc of drug discovery entails a multiparameter optimization problem spanning vast length scales. The key parameters range from solubility (angstroms) to protein–ligand binding (nanometers) to in vivo toxicity (meters). Through feature learning—instead of feature engineering—deep neural networks promise to outperform both traditional physics-based and knowledge-based machine learning models for predicting molecular properties pertinent to drug discovery. To this end, we present the PotentialNet family of graph convolutions. These models are specifically designed for and achieve state-of-the-art performance for protein–ligand binding affinity. We further validate these deep neural networks by setting new standards of performance in several ligand-based tasks. In parallel, we introduce a new metric, the Regression Enrichment Factor EFχ(R), to measure the early enrichment of computational models for chemical data. Finally, we introduce a cross-validation strategy based on structural homology clustering that can more accurately measure model generalizability, which crucially distinguishes the aims of machine learning for drug discovery from standard machine learning tasks.

The Absorption, Distribution, Metabolism, Elimination, and Toxicity (ADMET) properties of drug candidates are estimated to account for up to 50% of all clinical trial failures 1,2 . Predicting ADMET properties has therefore been of great interest to the cheminformatics and medicinal chemistry communities in recent decades. Traditional cheminformatics approaches, whether the learner is a random forest or a deep neural network, leverage fixed fingerprint feature representations of molecules. In contrast, in this paper, we learn the features most relevant to each chemical task at hand by representing each molecule explicitly as a graph, where each node is an atom and each edge is a bond. By applying graph convolutions to this explicit molecular representation, we achieve, to our knowledge, unprecedented accuracy in prediction of ADMET properties. By challenging our methodology with rigorous cross-validation procedures and prospective analyses, we show that deep featurization better enables molecular predictors to not only interpolate but also extrapolate to new regions of chemical space.

Empirical scoring functions based on either molecular force fields or cheminformatics descriptors are widely used, in conjunction with molecular docking, during the early stages of drug discovery to predict potency and binding affinity of a drug-like molecule to a given target. These models require expert-level knowledge of physical chemistry and biology to be encoded as hand-tuned parameters or features rather than allowing the underlying model to select features in a data-driven procedure. Here, we develop a general 3-dimensional spatial convolution operation for learning atomic-level chemical interactions directly from atomic coordinates and demonstrate its application to structure-based bioactivity prediction. The atomic convolutional neural network is trained to predict the experimentally determined binding affinity of a protein-ligand complex by direct calculation of the energy associated with the complex, protein, and ligand given the crystal structure of the binding pose. Non-covalent interactions present in the complex that are absent in the protein-ligand sub-structures are identified and the model learns the interaction strength associated with these features. We test our model by predicting the binding free energy of a subset of protein-ligand complexes found in the PDBBind dataset and compare with state-of-the-art cheminformatics and machine learning-based approaches. We find that all methods achieve experimental accuracy (less than 1 kcal/mol mean absolute error) and that atomic convolutional networks either outperform or perform competitively with the cheminformatics based methods. Unlike all previous protein-ligand prediction systems, atomic convolutional networks are end-to-end and fully-differentiable. They represent a new data-driven, physics-based deep learning model paradigm that offers a strong foundation for future improvements in structure-based bioactivity prediction.

No abstract

No abstract