Search for: All records

In this work, we investigate the question: do code-generating large language models know chemistry? Our results indicate, mostly yes. To evaluate this, we introduce an expandable framework for evaluating chemistry knowledge in these models, through prompting models to solve chemistry problems posed as coding tasks. To do so, we produce a benchmark set of problems, and evaluate these models based on correctness of code by automated testing and evaluation by experts. We find that recent LLMs are able to write correct code across a variety of topics in chemistry and their accuracy can be increased by 30 percentage points via prompt engineering strategies, like putting copyright notices at the top of files. Our dataset and evaluation tools are open source which can be contributed to or built upon by future researchers, and will serve as a community resource for evaluating the performance of new models as they emerge. We also describe some good practices for employing LLMs in chemistry. The general success of these models demonstrates that their impact on chemistry teaching and research is poised to be enormous. 

Inferring molecular structure from Nuclear Magnetic Resonance (NMR) measurements requires an accurate forward model that can predict chemical shifts from 3D structure. Current forward models are limited to specific molecules like proteins and state-of-the-art models are not differentiable. Thus they cannot be used with gradient methods like biased molecular dynamics. Here we use graph neural networks (GNNs) for NMR chemical shift prediction. Our GNN can model chemical shifts accurately and capture important phenomena like hydrogen bonding induced downfield shift between multiple proteins, secondary structure effects, and predict shifts of organic molecules. Previous empirical NMR models of protein NMR have relied on careful feature engineering with domain expertise. These GNNs are trained from data alone with no feature engineering yet are as accurate and can work on arbitrary molecular structures. The models are also efficient, able to compute one million chemical shifts in about 5 seconds. This work enables a new category of NMR models that have multiple interacting types of macromolecules. 

The selection of coarse-grained (CG) mapping operators is a critical step for CG molecular dynamics (MD) simulation. It is still an open question about what is optimal for this choice and there is a need for theory. The current state-of-the art method is mapping operators manually selected by experts. In this work, we demonstrate an automated approach by viewing this problem as supervised learning where we seek to reproduce the mapping operators produced by experts. We present a graph neural network based CG mapping predictor called Deep Supervised Graph Partitioning Model (DSGPM) that treats mapping operators as a graph segmentation problem. DSGPM is trained on a novel dataset, Human-annotated Mappings (HAM), consisting of 1180 molecules with expert annotated mapping operators. HAM can be used to facilitate further research in this area. Our model uses a novel metric learning objective to produce high-quality atomic features that are used in spectral clustering. The results show that the DSGPM outperforms state-of-the-art methods in the field of graph segmentation. Finally, we find that predicted CG mapping operators indeed result in good CG MD models when used in simulation.