Search

The lead optimization process in drug discovery campaigns is an arduous endeavour where the input of many medicinal chemists is weighed in order to reach a desired molecular property profile. Building the expertise to successfully drive such projects collaboratively is a very time-consuming process that typically spans many years within a chemist's career. In this work we aim to replicate this process by applying artificial intelligence learning-to-rank techniques on feedback that was obtained from 35 chemists at Novartis over the course of several months. We exemplify the usefulness of the learned proxies in routine tasks such as compound prioritization, motif rationalization, and biased \textit{de novo} drug design. Annotated response data is provided, and developed models and code made available through a permissive open-source license.

Explainable machine learning is increasingly used in drug discovery to help rationalize compound property predictions. Feature attribution techniques are popular choices within this toolbox that aim to identify which substructures of a molecule are responsible for a predicted property change. However, established molecular feature attribution methods have so far displayed low performance for popular deep learning algorithms such as graph neural networks (GNNs), particularly when compared with simpler modeling alternatives such as random forests coupled with atom masking. To mitigate this problem, in this work we present a modification to the regression objective of GNNs to specifically account for common core structures between pairs of molecules. The proposed approach shows higher accuracy on a recently-proposed explainability benchmark. We expect this methodology to be useful in drug discovery pipelines, and specifically in lead optimization efforts where specific chemical series of related compounds are investigated.

Many molecular design tasks benefit from fast and accurate calculations of quantum-mechanical (QM) properties. However, the computational cost of QM methods applied to drug-like molecules currently renders large-scale applications of quantum chemistry challenging. Aiming to mitigate this problem, we developed DelFTa, an open-source toolbox for the prediction of electronic properties of drug-like molecules at the density functional (DFT) level of theory, using Δ-machine-learning. Δ-Learning corrects the prediction error (Δ) of a fast but inaccurate property calculation. DelFTa employs state-of-the-art three-dimensional message-passing neural networks trained on a large dataset of QM properties. It provides access to a wide array of quantum observables on the molecular, atomic and bond levels by predicting approximations to DFT values from a low-cost semiempirical baseline. Δ-Learning outperformed its direct-learning counterpart for most of the considered QM endpoints. The results suggest that predictions for non-covalent intra- and intermolecular interactions can be extrapolated to larger biomolecular systems. The software is fully open-sourced and features documented command-line and Python APIs., Physical Chemistry Chemical Physics, 24 (18), ISSN:1463-9084, ISSN:1463-9076

Many molecular design tasks benefit from fast and accurate calculations of quantum-mechanical (QM) properties. However, the computational cost of QM methods applied to drug-like molecules currently renders large-scale applications of quantum chemistry challenging. Aiming to mitigate this problem, we developed DelFTa, an open-source toolbox for the prediction of electronic properties of drug-like molecules at the density functional (DFT) level of theory, using Δ-machine-learning. Δ-Learning corrects the prediction error (Δ) of a fast but inaccurate property calculation. DelFTa employs state-of-the-art three-dimensional message-passing neural networks trained on a large dataset of QM properties. It provides access to a wide array of quantum observables on the molecular, atomic and bond levels by predicting approximations to DFT values from a low-cost semiempirical baseline. Δ-Learning outperformed its direct-learning counterpart for most of the considered QM endpoints. The results suggest that predictions for non-covalent intra- and intermolecular interactions can be extrapolated to larger biomolecular systems. The software is fully open-sourced and features documented command-line and Python APIs.

Feature attribution techniques are popular choices within the explainable artificial intelligence toolbox, as they can help elucidate which parts of the provided inputs used by an underlying supervised-learning method are considered relevant for a specific prediction. In the context of molecular design, these approaches typically involve the coloring of molecular graphs, whose presentation to medicinal chemists can be useful for making a decision of which compounds to synthesize or prioritize. The consistency of the highlighted moieties alongside expert background knowledge is expected to contribute to the understanding of machine-learning models in drug design. Quantitative evaluation of such coloring approaches, however, has so far been limited to substructure identification tasks. We here present an approach that is based on maximum common substructure algorithms applied to experimentally-determined activity cliffs. Using the proposed benchmark, we found that molecule coloring approaches in conjunction with classical machine-learning models tend to outperform more modern, deep-learning-based alternatives. However, none of the tested feature attribution methods sufficiently and consistently generalized when confronted with unseen examples.

Certain molecular design tasks benefit from fast and accurate calculations of quantum-mechanical (QM) properties. However, the computational cost of QM methods applied to drug-like compounds currently makes large-scale applications of quantum chemistry challenging. In order to mitigate this problem, we developed DelFTa, an open-source toolbox for predicting small-molecule electronic properties at the density functional (DFT) level of theory, using the Δ-machine learning principle. DelFTa employs state-of-the-art E(3)-equivariant graph neural networks that were trained on the QMugs dataset of QM properties. It provides access to a wide array of quantum observables by predicting approximations to ωB97X-D/def2-SVP values from a GFN2-xTB semiempirical baseline. Δ-learning with DelFTa was shown to outperform direct DFT learning for most of the considered QM endpoints. The software is provided as open-source code with fully-documented command-line and Python APIs.

Certain molecular design tasks benefit from fast and accurate calculations of quantum-mechanical (QM) properties. However, the computational cost of QM methods applied to drug-like compounds currently makes large-scale applications of quantum chemistry challenging. In order to mitigate this problem, we developed DelFTa, an open-source toolbox for predicting small-molecule electronic properties at the density functional (DFT) level of theory, using the Δ-machine learning principle. DelFTa employs state-of-the-art E(3)-equivariant graph neural networks that were trained on the QMugs dataset of QM properties. It provides access to a wide array of quantum observables by predicting approximations to ωB97X-D/def2-SVP values from a GFN2-xTB semiempirical baseline. Δ-learning with DelFTa was shown to outperform direct DFT learning for most of the considered QM endpoints. The software is provided as open-source code with fully-documented command-line and Python APIs.

Graph neural networks are able to solve certain drug discovery tasks such as molecular property prediction and de novo molecule generation. However, these models are considered “black-box” and “hard-to-debug”. This study aimed to improve modeling transparency for rational molecular design by applying the integrated gradients explainable artificial intelligence (XAI) approach for graph neural network models. Models were trained for predicting plasma protein binding, hERG channel inhibition, passive permeability, and cytochrome P450 inhibition. The proposed methodology highlighted molecular features and structural elements that are in agreement with known pharmacophore motifs, correctly identified property cliffs, and provided insights into unspecific ligand–target interactions. The developed XAI approach is fully open-sourced and can be used by practitioners to train new models on other clinically relevant endpoints. © 2021 American Chemical Society, Journal of Chemical Information and Modeling, 61 (3), ISSN:1549-9596, ISSN:0095-2338, ISSN:1520-5142

While a plethora of different protein&ndash, ligand docking protocols have been developed over the past twenty years, their performances greatly depend on the provided input protein&ndash, ligand pair. In this study, we developed a machine-learning model that uses a combination of convolutional and fully connected neural networks for the task of predicting the performance of several popular docking protocols given a protein structure and a small compound. We also rigorously evaluated the performance of our model using a widely available database of protein&ndash, ligand complexes and different types of data splits. We further open-source all code related to this study so that potential users can make informed selections on which protocol is best suited for their particular protein&ndash, ligand pair.

Accurate in silico protein–ligand binding affinity prediction can substantially accelerate drug discovery pipelines by prioritizing compounds for experimental testing, a typically lengthy and costly process. Given the success of machine-learning and artificial intelligence approaches in areas such as computer vision and natural language processing in the last few years, there have been significant developments towards their application in structure-based potency prediction. In this chapter we summarize recent progress in this field, and we provide readers with a thorough introduction of the basic aspects to take into account when developing such models.

The capability to rank different potential drug molecules against a protein target for potency has always been a fundamental challenge in computational chemistry due to its importance in drug design. While several simulation-based methodologies exist, they are hard to use prospectively and thus predicting potency in lead optimization campaigns remains an open challenge. Here we present the first machine learning approach specifically tailored for ranking congeneric series based on deep 3D-convolutional neural networks. Furthermore we prove its effectiveness by blindly testing it on datasets provided by Janssen, Pfizer and Biogen totalling over 3246 ligands and 13 targets as well as several well-known openly available sets, representing one the largest evaluations ever performed. We also performed online learning simulations of lead optimization using the approach in a predictive manner obtaining significant advantage over experimental choice. We believe that the evaluation performed in this study is strong evidence of the usefulness of a modern deep learning model in lead optimization pipelines against more expensive simulation-based alternatives. The authors thank Acellera Ltd. for funding. G. D. F. acknowledges support from MINECO (BIO2014-53095-P), MICINN (PTQ-17-09079) and FEDER. This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 675451 (CompBioMed project).