The software package Chemprop implements the directed message-passing neural networks (D-MPNN) architecture and offers simple, easy, and fast access to machine-learned molecular properties. Compared to its initial version, we present a multitude of new Chemprop functionalities such as the support of multimolecule properties, reactions, atom/bond-level properties, and spectra.

Further, we incorporate various uncertainty quantification and calibration methods along with related metrics as well as pretraining and transfer learning workflows, improved hyperparameter optimization, and other customization options concerning loss functions or atom/bond features.

We benchmark D-MPNN models trained using Chemprop with the new reaction, atom-level, and spectra functionality on a variety of property prediction data sets, including MoleculeNet and SAMPL, and observe state-of-the-art performance on the prediction of water-octanol partition coefficients, reaction barrier heights, atomic partial charges, and absorption spectra.

We present a method for creating RDKit parsable SMILES for transition metal complexes (TMCs) based on xyz-coordinates and overall charge of the complex. This can be viewed as an extension to the program xyz2mol that does the same for organic molecules. The only dependency is RDKit, which makes it widely applicable. One thing that has been lacking when it comes to generating SMILES from structure for TMCs is an existing SMILES dataset to compare with.

In this work, phosphorescent materials are represented as strings, molecular graphs, and point clouds, which are employed by language models, two-dimensional graph, and three-dimensional graph neural networks. In addition, more than 200 000 molecules with simulated properties highly relevant to experimental properties are used for pretraining the DL models.

Our work shows high performance in the prediction of five experimental properties that are importantly considered when commercializing OLED devices. This means that faster material discovery for OLEDs can be achieved through DL models that are trained with simulation information that is highly correlated with experimental properties.

In this study, we present a comprehensive benchmark of carbon shielding anisotropies based on coupled cluster reference tensors taken from the NS372 benchmark data set.

Additionally, we investigate the representation of the DFT-predicted shielding tensors, such as the eigenvalues and eigenvectors. Moreover, we evaluated how various DFT methods influence the discrimination of possible relative configurations using recently published ΔΔRCSA data for a set of structurally diverse natural products.

Our findings demonstrate that accurate interpretation of RCSAs for configurational and conformational analysis is possible with semilocal DFT methods, which also reduce computational demands compared to hybrid functionals such as the commonly used B3LYP.

Briefly summarized are our and others’ experiences with the AI/ML applications that currently have the greatest impact on our work.

Utilizing a chain of advanced large language models (LLMs), we developed a systematic approach to extract and organize MOF data into a structured format.

Our methodology successfully compiled information from more than 40,000 research articles, creating a comprehensive and ready-to-use data set. Specifically, data regarding MOF synthesis conditions and properties were extracted from both tables and text and then analyzed. Subsequently, we utilized the curated database to analyze the relationships between synthesis conditions, properties, and structure.

We applied a classifier method to predict palladium catalysts for the formation of nonalternating polyketones via the copolymerization of CO and ethylene; current examples are limited to using phosphine sulfonate and diphosphazane monoxide supporting ligands.

With the reported workflow, we discovered two new classes of palladium complexes capable of achieving the synthesis of nonalternating polyketones with a lower CO content than those made by known palladium catalysts.

Our results show that we doubled the number of classes of palladium compounds that can catalyze the formation of this type of polymer. We envision that this methodology can be applied to accelerate catalyst discovery when selectivity is an important outcome.

This review highlights LLM capabilities in these domains and their potential to accelerate scientific discovery through automation. We also review LLM-based autonomous agents: LLMs with a broader set of tools to interact with their surrounding environment. These agents perform diverse tasks such as paper scraping, interfacing with automated laboratories, and synthesis planning.

As agents are an emerging topic, we extend the scope of our review of agents beyond chemistry and discuss across any scientific domains.

This review covers the recent history, current capabilities, and design of LLMs and autonomous agents, addressing specific challenges, opportunities, and future directions in chemistry.

Herein, we review the recent advancements in AI applications for retrosynthesis prediction by summarizing related techniques and the landscape of current representative retrosynthesis models and propose feasible solutions to tackle existing problems and outline future directions in this field.

The development of machine learning models to predict the regioselectivity of C(sp3)–H functionalization reactions is reported. A data set for dioxirane oxidations was curated from the literature and used to generate a model to predict the regioselectivity of C–H oxidation.

To assess whether smaller, intentionally designed data sets could provide accuracy on complex targets, a series of acquisition functions were developed to select the most informative molecules for the specific target.

Finally, the workflow was experimentally validated on five complex substrates and shown to be applicable to predicting the regioselectivity of arene C–H radical borylation.

AI-predicted routes are typically compared to experimental syntheses to check for an exact match among the top-ranked predictions (top-N accuracy).

This method is ideal for the evaluation of retrosynthetic algorithms on large datasets (>10^6 routes), but it cannot assess a degree of similarity between routes, which would be desirable for small datasets (<10^2 routes).

Here, we present a simple method to calculate a similarity score between any two synthetic routes to a given molecule.

The score is based on two concepts: which bonds are formed during the synthesis; and how the atoms of the final compound are grouped together throughout the synthesis. As a result, the similarity score overlaps well with chemists' intuition and provides a finer assessment of prediction accuracy.

In this work, we propose a general method for combining molecular descriptors with representation learning using the so-called expectation maximization algorithm from the probabilistic machine-learning literature, which uses uncertainty estimates to trade off between the two approaches.

The proposed hybrid model exploits chemical structure information using graph neural networks, but it automatically detects cases where structure-based predictions are unreliable, in which case it corrects them by representation-learning based predictions that can better specialize to unusual cases. The effectiveness of the proposed method is demonstrated using the prediction of activity coefficients in binary mixtures as an example.

The results are compelling, as the method significantly improves predictive accuracy over the current state of the art, showcasing its potential to advance the prediction of physico-chemical properties in general.

Accurate prediction of thermodynamic solubility by machine learning remains a challenge.

We investigated over 20 years of published solubility datasets and models, highlighting overlooked datasets and the overlaps between popular sets.

We benchmarked recently published models on a novel curated solubility dataset and report poor performances. We also propose a workflow to cure aqueous solubility data aiming at producing useful models for bench chemist.

Our results demonstrate that some state-of-the-art models are not ready for public usage because they lack a well-defined applicability domain and overlook historical data sources.

We report the impact of factors influencing the utility of the models: interlaboratory standard deviation, ionic state of the solute and data sources. The herein obtained models, and quality-assessed datasets are publicly available.