In recent years, there have been considerable academic and industrial research efforts to develop novel generative models for high-performing, small molecules. Traditional, rules-based algorithms such as genetic algorithms [Jensen, Chem. Sci., 2019, 12, 3567-3572] have, however, been shown to rival deep learning approaches in terms of both efficiency and potency. In previous work, we showed that the addition of a quality-diversity archive to a genetic algorithm resolves stagnation issues and substantially increases search efficiency [Verhellen, Chem. Sci., 2020, 42, 11485-11491]. In this work, we expand on these insights and leverage the availability of bespoke kernels for small molecules [Griffiths, Adv. Neural. Inf. Process. Syst., 2024, 36] to integrate Bayesian optimisation into the quality-diversity process. This novel generative model, which we call Bayesian Illumination, produces a larger diversity of high-performing molecules than standard quality-diversity optimisation methods. In addition, we show that Bayesian Illumination further improves search efficiency com- pared to previous generative models for small molecules, including deep learning approaches, genetic algorithms, and standard quality-diversity methods.
A set of five compounds containing peroxo, superoxo or bis-mu-oxo moieties has been studied in the gas phase using CCSD(T)/aug-cc-pVQZ, also in combination with Goodson’s continued fraction approach. The corresponding analytical frequencies corroborate assignments of bands from experiments, and thus provide a consistent set of reference data that can be used for benchmarking a range of density functional approximations. A total of 100 density functionals have been checked for the general bond lengths, the specific peroxo/superoxo bond lengths, angles, and vibrational frequencies. There is not one density functional that performs equally well for all of these properties, not even within one class of density functionals.
The high chemical stability of SU-8 makes it irreplaceable for a wide range of applica- tions, most notably as a lithography photoresist for micro and nanotechnology. This advantage becomes a problem when there is a need to remove SU-8 from the fabricated devices. Researchers have been struggling for two decades with this problem, and al- though a number of partial solutions have been found, this difficulty has limited the applications of SU-8. Here we demonstrate a fast, reproducible, and comparatively gen- tle method to chemically remove SU-8 photoresist. An ether cleavage mechanism for the observed reaction is proposed, and the hypothesis is tested with ab initio quantum chem- ical calculations. Also described are a polymer-metal adhesion treatment, and a comple- mentary removal method, based on atomic hydrogen inductively coupled plasma.
While there is a great deal of interest in methods aimed at explaining machine learning predictions of chemical properties, it is difficult to quantitatively benchmark such methods, especially for regression tasks. We show that the Crippen logP model (J. Chem. Inf. Comput. Sci. 1999, 39, 868) provides an excellent benchmark for atomic attribution/heatmap approaches, especially if the ground truth heatmaps can be adjusted to reflect the molecular representation. The “atom attribution from finger prints”-method developed by Riniker and Landrum (J. Chem. Inf. Comput. Sci. 2013, 5, 43) gives atomic attribution heatmaps that are in reasonable agreement with the atomic contribution heatmaps of the Crippen logP model for most molecules, with average heatmap overlaps of up to 0.54. The agreement is increased significantly (to 0.75) when the atomic contributions are adjusted to match the fact that the molecular representation is fragment-based rather than atom-based (the finger print-adapted (FPA) ground truth vector). Most heatmaps and the corresponding FPA overlaps are relatively insensitive to the training set size and the results are close to converged for a training set size of 1000 molecules, although for molecules with low overlap some heatmaps change significantly. Using the “remove an atom” approach for graph convolutional neural networks (GCNNs) suggested by Matveieva and Polishchuk (J. Cheminform. 2021, 13, 41) we find an average heatmap overlap of 0.47 for the atomic contribution heatmaps of the Crippen logP model. Like the simpler attribution benchmarks for classification tasks that have come before it, this work sets the bar for regression tasks.
As a new approach to efficiently describe correlation effects in the relativistic quantum world we propose to consider reduced density matrix functional theory, where the key quantity is the firstorder reduced density matrix (1-RDM). In this work, we first introduce the theoretical foundations to extend the applicability of this theory to the relativistic domain. Then, using the so-called no-pair (np) approximation, we arrive at an approximate treatment of the relativistic effects by focusing on electronic wavefunctions and neglecting explicit contributions from positrons. Within the np approximation the theory becomes similar to the nonrelativistic case, with as unknown only the functional that describes the electron-electron interactions in terms of the 1-RDM. This requires the construction of functional approximations, and we therefore also present the relativistic versions of some common RDMFT approximations that are used in the nonrelativistic context and discuss their properties.
