Riniker lab @ ETHZ

Efficient Multistate Free-Energy Calculations with QM/MM Accuracy Using Replica-Exchange Enveloping Distribution Sampling Calculating free-energy differences using molecular dynamics (MD) simulations is an important task in computational chemistry. In practice, the accuracy of the results is limited by model approximatio...

Our newest publication explores using QM/MM together with RE-EDS to do efficient free-energy calculations, enabling increased accuracy and allowing highly polarized systems to be studied.
#chemsky #compchem
doi.org/10.1021/acs....

Neural Network Potential with Multiresolution Approach Enables Accurate Prediction of Reaction Free Energies in Solution We present the design and implementation of a novel neural network potential (NNP) and its combination with an electrostatic embedding scheme, commonly used within the context of hybrid quantum-mechan...

Our new paper in JACS presents the application of a new neural network potential, based on our anisotropic message passing approach, to do QM/MM MD simulations. We achieve chemical accuracy on a few quite different types of chemistry.

doi.org/10.1021/jacs...
#chemsky #compchem #opensource

Hi Marwin. Could you please add us too?
Thanks!

Since it's flexible and can be directly integrated into Python code, we think it's a very useful tool for computational science.

lwreg: A Lightweight System for Chemical Registration and Data Storage Here, we present lwreg, a lightweight, yet flexible chemical registration system supporting the capture of both two-dimensional molecular structures (topologies) and three-dimensional conformers. lwre...

Our newest preprint describes lwreg, a lightweight system for chemical registration. lwreg has a Python API and makes it easy to store the compound structures you use in your work and the experimental data you generate about them.

chemrxiv.org/engage/chemr...

Our paper introducing an implicit solvation model for organic molecule in water based on a graph neural network has just appeared:
pubs.rsc.org/en/content/a...
#chemsky

Probing the Stability of a β-Hairpin Scaffold after Desolvation Probing the structural characteristics of biomolecular ions in the gas phase following native mass spectrometry (nMS) is of great interest, because noncovalent interactions, and thus native fold featu...

The most recent paper from our ongoing collaboration with the Zenobi group looks at the impact of desolvation on the stability of beta-hairpin structures.
#chemsky

pubs.acs.org/doi/10.1021/...

General Graph Neural Network Based Implicit Solvation Model for Organic Molecules in Water The dynamical behavior of small molecules in their environment can be studied with molecular dynamics (MD) simulations to gain deeper insight on an atomic level and thus complement and rationalize the...

Our newest preprint, introducing a new type of implicit solvation model based on a graph neural network, is now up: chemrxiv.org/engage/chemr...

Unraveling motion in proteins by combining NMR relaxometry and molecular dynamics simulations: A case study on ubiquitin Nuclear magnetic resonance (NMR) relaxation experiments shine light onto the dynamics of molecular systems in the picosecond to millisecond timescales. As these

In our newest paper we look at using five different computational methods applied to the results of MD simulations and NMR relaxation experiments in order to better understand protein motions.
doi.org/10.1063/5.01...

Our most recent paper just appeared in JCIM. The title of this one pretty much tells the story: when you assemble a data set by combining data from different literature assays, there is a very good chance that the resulting data contains a lot of noise.
pubs.acs.org/doi/10.1021/...

Combining IC50 or Ki Values From Different Sources is a Source of Significant Noise As part of the ongoing quest to find or construct large data sets for use in validating new machine learning (ML) approaches for bioactivity prediction, it has become distressingly common for research...

We have a new preprint out which looks at the amount of noise introduced into a data set when we combine data from different ChEMBL assays.
doi.org/10.26434/che...

SIMPD: an algorithm for generating simulated time splits for validating machine learning approaches ... Time-split cross-validation is broadly recognized as the gold standard for validating predictive models intended for use in medicinal chemistry projects. Unfortunately this type of data is not broadly...

Our paper introducing SIMPD is now out. SIMPD is an algorithm for creating training/test sets for molecular #machinelearning based on an analysis of a large number of real-world medchem projects.
link.springer.com/article/10.1...
#opensource code and data are in github.
github.com/rinikerlab/m...

Unraveling Motion in Proteins by Combining NMR Relaxometry and Molecular Dynamics Simulations: A Cas... Nuclear magnetic resonance (NMR) relaxation experiments shine light onto the dynamics of molecular systems in the picosecond to nanosecond timescales. As these methods cannot provide an atomically res...

Our most recent preprint describes research done together with the Ferrage group at the Sorbonne in Paris to apply #MolecularDynamics and #NMR to understand protein motions in solution.
chemrxiv.org/engage/chemr...

Hybrid classical/machine-learning force fields for the accurate description of molecular condensed-p... Electronic structure methods offer in principle accurate predictions of molecular properties, however, their applicability is limited by computational costs. Empirical methods are cheaper, but come wi...

Our most recent publication describes a hybrid classical/machine-learning forcefield we've developed for condensed-phase systems.
As usual, it's #opensource, #opendata, and #openaccess
pubs.rsc.org/en/content/a...

Our paper introducing DASH, an efficient approach for assigning partial charges to atoms in molecules is now out. The method uses a hierarchy created from attention values from a GNN trained on QM data.
It's #opensource, #opendata, and #openaccess
pubs.acs.org/doi/10.1021/...