Hot off the press. We improved our IR spectra alignment algorithm that helps you determining the relative stereochemistry of organic molecules. It compares theoretical and experimental IR spectra of all isomers and finds the best match. Check it out here: pubs.rsc.org/en/content/art…

Preprint alert. Interested in water molecules that affect ligand binding and affinity to protein pockets? We analyzed their energetic contributions as well as arising solvation correlation effects of water networks. #ComputationalChemistry #DrugDesign chemrxiv.org/engage/chemrxi…

"Regularized by Physics": our new paper introduces a method for using graph neural networks to parameterize force-field potentials. pubs.acs.org/doi/10.1021/ac…

Great teamwork and discussions with academic and industry colleagues lead to this perspective looking at machine learning for small-molecule drug discovery. sciencedirect.com/science/articl…

Our newest paper introduces enu, an #opensource tool for enumerating all isomers (including stereoisomers!) for a given molecular formula. jcheminf.biomedcentral.com/articles/10.11… You can find the source code for enu on GitHub: github.com/csms-ethz/Comb…

We are very happy to be starting this multi-year project with our collaborators at Novartis Science.

We've got a new #preprint up describing an approach for simulating an implicit solvent for MD using transferable graph neural networks. All the code is, of course, #opensource. arxiv.org/abs/2302.09321

Escape from the curse of dimensionality: Our newest preprint introduces an energy-based clustering algorithm that allows you to avoid common problems with estimated densities. The code is, of course, #opensource arxiv.org/abs/2303.01104 github.com/rinikerlab/Ene…

Our new #preprint introduces SIMPD: an algorithm for creating training/test sets for molecular #machinelearning based on an analysis of a large number of real-world medchem projects. #opensource code and data are in github. chemrxiv.org/engage/chemrxi…

Interested in water in ligand binding sites? Our newest paper applies RE-EDS to study the thermodynamics of water networks in protein cavities. pubs.acs.org/doi/10.1021/ac…

In our most recent publication we use a graph neural network to develop a model for implicit solvation in molecular dynamics simulations of peptides. The code is open source and the model is freely downloadable. pubs.aip.org/aip/jcp/articl…

Our second paper that came out today describes Anna Sophia Kamenik's collaboration with the Zenobi group combining experimental and computational techniques to help understand the structures of peptides in the gas phase. nature.com/articles/s4146…

Our most recent preprint shares an efficient approach for assigning partial charges to atoms in molecules. The method uses a hierarchy created from attention values from a GNN trained on QM data. arxiv.org/abs/2305.15981 It's all #opensource and #opendata, of course.

The next paper from our collaboration with the Zenobi Lab is now out. Here we applied #MolecularDynamics to understand the trends observed in #FRET experiments. pubs.acs.org/doi/10.1021/ac…

We're really happy that our most recent publication in The Journal of Chemical Physics, which introduces a new energy-based clustering algorithm, has been featured by the editors will be the cover image for the next issue. pubs.aip.org/aip/jcp/articl…

We are really happy to announce that Stephanie Linker's paper on membrane permeability of cyclic peptides has won the Royal Society of Chemistry's Outstanding Student Paper award in computational and theoretical chemistry for 2022!

Our latest preprint introduces and validates a new hybrid classical/ML force field for working with condensed phase systems. It's now up on arXiv: arxiv.org/abs/2308.08984

We're happy to be the #ChemSciPicks of week from Chemical Science