The QM9 dataset has become the golden standard for Machine Learning (ML) predictions of various chemical properties. QM9 is based on the GDB, which is a combinatorial exploration of the chemical space. ML molecular predictions have been recently published with an accuracy on par with Density Functional Theory calculations. Such ML models need to be tested and generalized on real data. PC9, a new QM9 equivalent dataset (only H, C, N, O and F and up to 9 “heavy” atoms) of the PubChemQC project is presented in this article. A statistical study of bonding distances and chemical functions shows that this new dataset encompasses more chemical diversity. Kernel Ridge Regression, Elastic Net and the Neural Network model provided by SchNet have been used on both datasets. The overall accuracy in energy prediction is higher for the QM9 subset. However, a model trained on PC9 shows a stronger ability to predict energies of the other dataset.

The objective of this work is to design a molecular generator capable of exploring known as well as unfamiliar areas of the chemical space. Our method must be flexible to adapt to very different problems. Therefore, it has to be able to work with or without the influence of prior data and knowledge. Moreover, regardless of the success, it should be as interpretable as possible to allow for diagnosis and improvement. We propose here a new open source generation method using an evolutionary algorithm to sequentially build molecular graphs. It is independent of starting data and can generate totally unseen compounds. To be able to search a large part of the chemical space, we define an original set of 7 generic mutations close to the atomic level. Our method achieves excellent performances and even records on the QED, penalised logP, SAscore, CLscore as well as the set of goal-directed functions defined in GuacaMol. To demonstrate its flexibility, we tackle a very different objective issued from the organic molecular materials domain. We show that EvoMol can generate sets of optimised molecules having high energy HOMO or low energy LUMO, starting only from methane. We can also set constraints on a synthesizability score and structural features. Finally, the interpretability of EvoMol allows for the visualisation of its exploration process as a chemically relevant tree.

This paper reports SVR (Support Vector Regression) and GTM (Generative Topographic Mapping) modeling of three kinetic properties of cycloaddition reactions: rate constant (logk), activation energy (Ea) and pre-exponential factor (logA). A data set of 1849 reactions, comprising (4+2), (3+2) and (2+2) cycloadditions (CA) were studied in different solvents and at different temperatures. The reactions were encoded by the ISIDA fragment descriptors generated for Condensed Graph of Reaction (CGR). For a given reaction, a CGR condenses structures of all the reactants and products into one single molecular graph, described both by conventional chemical bonds and "dynamical" bonds characterizing chemical transformations. Different scenarios of logk assessment were exploited: direct modeling, application of the Arrhenius equation and temperature-scaled GTM landscapes. The logk models with optimal cross-validated statistics (Q =0.78-0.94 RMSE=0.45-0.86) have been challenged to predict rates for the external test set of 200 reactions, comprising both reactions that were not present in the training set, and training set transformations performed under different reaction conditions. The models are freely available on our web-server: http://cimm.kpfu.ru/models.