New Publication on LSER Descriptors Released

We are excited to announce our latest publication, “Prediction of solute descriptors for linear solvation energy relationships using k-nearest neighbours, group contributions, and graph-convolutional neural networks” (ACS Environ. Au, 2026, https://doi.org/10.1021/acsenvironau.6c00063). In this study, we demonstrate that our graph-convolutional neural network (GNN) approach performs the best, significantly outperforming the other models. Compared to our previous model variants, we were able to greatly enhance predictive performance and substantially reduce prediction errors.

The best part for our users: These powerful new models are already implemented in the LSER Toolbox and now serve as the foundation for calculating descriptors for all LSER equations within our Pauly software. With this update, we ensure that our users are equipped with the best possible prediction models for LSER descriptors, directly accessible within our software.