The development of force-responsive molecules called mechanophores is a central component of the field of polymer mechanochemistry. Mechanophores enable the design and fabrication of polymers for a variety of applications ranging from sensing to molecular release and self-healing materials. Nevertheless, an insufficient understanding of structure−activity relationships limits experimental development, and thus computation is necessary to guide the structural design of mechanophores. The constrained geometries simulate external force (CoGEF) method is a highly accessible and straightforward computational technique that simulates the effect of mechanical force on a molecule and enables the prediction of mechanochemical reactivity. Here, we use the CoGEF method to systematically evaluate every covalent mechanophore reported to date and compare the predicted mechanochemical reactivity to experimental results. Molecules that are mechanochemically inactive are also studied as negative controls. In general, mechanochemical reactions predicted with the CoGEF method at the common B3LYP/6-31G* level of density functional theory are in excellent agreement with reactivity determined experimentally. Moreover, bond rupture forces obtained from CoGEF calculations are compared to experimentally measured forces and demonstrated to be reliable indicators of mechanochemical activity. This investigation validates the CoGEF method as a powerful tool for predicting mechanochemical reactivity, enabling its widespread adoption to support the developing field of polymer mechanochemistry. Secondarily, this study provides a contemporary catalog of over 100 mechanophores developed to date.
We demonstrate that fast and accurate linear force fields can be built for molecules using the atomic cluster expansion (ACE) framework. The ACE models parametrize the potential energy surface in terms of body-ordered symmetric polynomials making the functional form reminiscent of traditional molecular mechanics force fields. We show that the four- or five-body ACE force fields improve on the accuracy of the empirical force fields by up to a factor of 10, reaching the accuracy typical of recently proposed machine-learning-based approaches. We not only show state of the art accuracy and speed on the widely used MD17 and ISO17 benchmark data sets, but we also go beyond RMSE by comparing a number of ML and empirical force fields to ACE on more important tasks such as normal-mode prediction, high-temperature molecular dynamics, dihedral torsional profile prediction, and even bond breaking. We also demonstrate the smoothness, transferability, and extrapolation capabilities of ACE on a new challenging benchmark data set comprised of a potential energy surface of a flexible druglike molecule.
Organic synthesis remains a major challenge in drug discovery. Although a plethora of machine learning models have been proposed as solutions in the literature, they suffer from being opaque black-boxes. It is neither clear if the models are making correct predictions because they inferred the salient chemistry, nor is it clear which training data they are relying on to reach a prediction. This opaqueness hinders both model developers and users. In this paper, we quantitatively interpret the Molecular Transformer, the state-of-the-art model for reaction prediction. We develop a framework to attribute predicted reaction outcomes both to specific parts of reactants, and to reactions in the training set. Furthermore, we demonstrate how to retrieve evidence for predicted reaction outcomes, and understand counterintuitive predictions by scrutinising the data. Additionally, we identify Clever Hans predictions where the correct prediction is reached for the wrong reason due to dataset bias. We present a new debiased dataset that provides a more realistic assessment of model performance, which we propose as the new standard benchmark for comparing reaction prediction models.
The rapid progress of machine learning interatomic potentials over the past couple of years produced a number of new architectures. Particularly notable among these are the Atomic Cluster Expansion (ACE), which unified many of the earlier ideas around atom density-based descriptors, and Neural Equivariant Interatomic Potentials (NequIP), a message passing neural network with equivariant features that showed state of the art accuracy. In this work, we construct a mathematical framework that unifies these models: ACE is generalised so that it can be recast as one layer of a multi-layer architecture. From another point of view, the linearised version of NequIP is understood as a particular sparsification of a much larger polynomial model. Our framework also provides a practical tool for systematically probing different choices in the unified design space. We demonstrate this by an ablation study of NequIP via a set of experiments looking at in-and out-of-domain accuracy and smooth extrapolation very far from the training data, and shed some light on which design choices are critical for achieving high accuracy. Finally, we present BOTNet (Body-Ordered-Tensor-Network), a much-simplified version of NequIP, which has an interpretable architecture and maintains accuracy on benchmark datasets.
We demonstrate that accurate linear force fields can be built using the Atomic Cluster Expansion (ACE) framework for molecules. Our model is built from body ordered symmetric polynomials which makes it a natural extension of traditional molecular mechanics force fields, and the large number of free parameters allows sufficient flexibility that it reaches the accuracy typical of recently proposed machine learning based approaches. We test our model on the MD17 and ISO17 data sets and also on a larger, more flexible molecule, and compare to leading machine learning models as well as refitted empirical force fields. We show that the linear body ordered ACE model has excellent transferability for properties beyond raw energy and force RMSE, both for molecular dynamics at different temperatures and for configurations very far from the training set including dihedral scans and even bond breaking.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.