New paper in the JCP Becke special issue. We show that Δ-learning can improve the transferability of MLIP potentials by combining an ANI-style MLIP with a DFTB3 baseline. The approach improves transition-state energetics and long-range interactions beyond the ML cutoff. doi.org/10.1063/5.03...

Δ-learning for transferable machine learning interatomic potentials

Machine-learning interatomic potentials (MLIPs) trained by directly learning the total interatomic interaction energies can suffer from limited transferability,

doi.org

🚨 Can AI reliably handle real-world, multi-step MD workflows comp chemists face every day? With Nithishwer Mouroug Anand and colleagues in the Biggin Lab @philbiggin.bsky.social @oxfordbiochemistry.bsky.social, we built MDArena from problems in active research projects. arxiv.org/abs/2608.02642

MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows

Accelerating scientific discovery is among the most consequential applications of AI, and computational biomolecular simulation stands out as a particularly promising target within this broader effort...

arxiv.org

Latest preprint where we pooled our sim experiences to create "MDArena" for AI agents trying to automate MD simulation and analysis. Great work led by Nithishwer Mouroug Anand and @weitse-hsu.bsky.social and thanks to all lab members! @oxfordbiochemistry.bsky.social arxiv.org/abs/2608.02642

MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows

Accelerating scientific discovery is among the most consequential applications of AI, and computational biomolecular simulation stands out as a particularly promising target within this broader effort...

arxiv.org

📣 ❗2nd Preprint of the Summer ❗ 📣 "Kinase inhibitors can change protonation or tautomeric state upon binding." LInk: doi.org/10.64898/202... We find that kinase-inhibitors can change in net-charge and tautomer state upon binding (from solution-state to kinase-bound), affecting med-chem choices!

Now out as a perspective in PLOS Biology. Where next for structural bioinformatics? journals.plos.org/plosbiology/...

Joe Greener@jgreener64.bsky.social · 11mo ago

I wrote a blog post about the future of structural bioinformatics. Where to go after AlphaFold? How do we avoid the field becoming a load of half-baked LLMs? Let me know what you think. jgreener64.github.io/posts/struct...

**Please repost** If you're enjoying Paper Skygest -- our personalized feed of academic content on Bluesky -- we'd appreciate you reposting this! We’ve found that the most effective way for us to reach new users and communities is through users sharing it with their network

In structure-based drug design, ligand conformational strain helps answer a key question: is this bound pose chemically plausible? But the usual gas-phase shortcut can misestimate strain, especially for charged and highly polar ligands. #compchem

AF-CALVADOS is now published doi.org/10.1002/pro.... We combine AlphaFold and CALVADOS to simulate flexible multidomain proteins at scale: — Ensembles of >12000 full-length human proteins — Comparison of IDRs alone and I n context for >1500 TFs @sobuelow.bsky.social @kejohansson.bsky.social

New preprint! Do all kinase domains have similar activation mechanisms? We steer AlphaFold2 using "transfer seeding" of known kinases to sample 8.3 ms of MD simulations via Folding@home. We find unique pathways that are kinase-specific and depend on different elements! Link: doi.org/10.64898/202...

Interested in training bespoke molecular mechanics force fields quickly? Give our package a go github.com/cole-group/p... !

Cole Group@colegroupncl.bsky.social · 3mo ago

📢 New preprint: "Fast training of bespoke SMIRNOFF-format molecular mechanics force fields using machine learning potentials", by @finlayclark.bsky.social et al. chemrxiv.org/doi/full/10....

This study has been two years in the making and the results have already been widely discussed and compared against, but the paper is finally out in JCIM today! pubs.acs.org/doi/10.1021/...

Large-Scale Collaborative Assessment of Binding Free Energy Calculations for Drug Discovery Using OpenFE

Accurately measuring compound binding affinities is key to driving the pharmaceutical development process. Rigorous physics-based in silico approaches, particularly alchemical free energy methods, hav...

pubs.acs.org

Our new FEP sampling engine GrandFEP is now on GitHub and Chemrxvi. We implemented GCMC, Water-Swap MC, REST2, and terminal-flip MC in OpenMM. Github github.com/deGrootLab/G... Chemrxiv doi.org/10.26434/che...

RSC CICAG Chemical Structure Representations Meeting 2026 Burlington House, London, UK Wednesday 8th April registrations.hg3conferences.co.uk/hg3/frontend... Fabulous line up of speakers.

Check out our pre-print, where we train a protein and small molecule force field from scratch with a graph neural network. We show comparable performance to existing, manually-tuned force fields on a range of tasks including binding free energy prediction. (1/4) arxiv.org/abs/2603.16770

Training a force field for proteins and small molecules from scratch

Force fields for molecular dynamics are usually developed manually, limiting their transferability and making systematic exploration of functional forms challenging. We developed a graph neural networ...

arxiv.org

Modern molecular modeling needs a new mode of software development. Consortia like Open Free Energy build shared tools and release code under open licenses. @omsf.io aligns incentives across stakeholders, enabling an ecosystems that elevates the entire community. pubs.acs.org/doi/10.1021/...

The Open Molecular Software Foundation (OMSF) and the Growing Role of Open Source Software in Molecular Modeling

The increasing importance and predictive power of modern molecular modeling, driven by physics- and machine-learning-based methods, necessitates a new collaborative architecture to replace the isolate...

pubs.acs.org

Many #machinelearning potentials are built (or understood) in terms of "atomic cluster expansions" that link directly to a body-ordered energy decomposition that can be computed explicitly with a sequence of electronic structure calculations. But what kind of expansion do they learn in practice? A🧵

"In contrast to previous attempts at iterative optimization, we employ a validation set to determine convergence. Using a validation set circumvents problems with parameter convergence and flags when overfitting occurs" pubs.acs.org/doi/full/10....

Robust and Automated Force Field Parameterization Using Validation Sets and Active Learning

Molecular mechanics force fields enable atomistic simulations of complex systems that are too large for a quantum mechanical treatment. Simulation accuracy depends on the parameters employed in the fo...

pubs.acs.org

Now out in JACS! 🎉 : "Computing Solvation Free Energies of Small Molecules with Experimental Accuracy"! It's been a pleasure to collaborate on this with Harry Moore (@jhmchem.bsky.social) & Gábor Csányi pubs.acs.org/doi/10.1021/...

📢 Can AI-Predicted Complexes Teach Machine Learning to Compute Drug Binding Affinity? In our recent JCIM work, we tested whether co-folding models can be used for data augmentation for training ML-based scoring functions (SFs). We asked 3 simple but critical questions. 👇 (1/6)

Ever wanted to run MD simulations of entire proteins in water with DFT accuracy? Meet AMPv3-BMS25, the latest iteration of our AMP multiscale neural network potential by @rinikerlab.bsky.social Read more in the preprints: doi.org/10.26434/che... doi.org/10.26434/che...

We’re pleased to announce the full release of the Sage 2.3.0 force field! This is identical to the previous release candidate Sage 2.3.0rc2. Sage 2.3.0 is the first OpenFF force field to use the AshGC neural network charge model. github.com/openforcefie... #compchem

Release Sage 2.3.0 · openforcefield/openff-forcefields

This release adds openff-2.3.0.offxml and openff_unconstrained-2.3.0.offxml. Sage 2.3.0 is the first OpenFF force field to use the AshGC neural network charge model to assign charges. Both vdW para...

github.com

OpenFE is ready for production! chemrxiv.org/engage/chemr... In collaboration with our industry partners, we ran benchmarking simulations of our hybrid-topology RBFE protocol on a large collection of both public and private protein-ligand binding datasets. #compchem

Large-scale collaborative assessment of binding free energy calculations for drug discovery using OpenFE

Accurately measuring compound binding affinities is key to driving the pharmaceutical development process. Rigorous physics-based in silico approaches, particularly alchemical free energy methods, hav...

chemrxiv.org

New preprint! How many crystal structures do you need to trust your docking results? It turns out, if you know how your main scaffold binds in the pocket, not that many! Check out this work from recently-defended lab member @keysandcompounds.bsky.social On BioRxiv: www.biorxiv.org/content/10.1...

Congrats to Wei-Tse Hsu on our latest paper with Aniket Magarkar @boehringerglobal.bsky.social where we look at how well co-folding models might be used in the context of data-augmentation for affinity prediction models. pubs.acs.org/doi/10.1021/...

Can AI-Predicted Complexes Teach Machine Learning to Compute Drug Binding Affinity?

We evaluate the feasibility of using co-folding models for synthetic data augmentation in training machine learning-based scoring functions (MLSFs) for binding affinity prediction. Our results show th...

pubs.acs.org

📢 Looking for a PhD in computational drug discovery? Check out this funded opportunity with @agnesnoy.bsky.social at York, in collaboration with researchers at Newcastle, Oxford & Inspiralis! ⬇️

Agnes Noy@agnesnoy.bsky.social · 10mo ago

Would you like to fight against antimicrobial resistance (AMR) developing antibiotics? Apply to the PhD opportunity below and come to York. Collaboration with colegroupncl.bsky.social, @dghilarov.bsky.social and Inspiralis Ltd. Fully funded via DiMeN DTP 💡 www.findaphd.com/phds/project...

📢 We have a fully funded PhD studentship available for Oct 2026 start on "Training force fields for computer-aided drug design with machine learning", in collaboration with Ioan Magdau and SandboxAQ. Full details and how to apply: www.ncl.ac.uk/postgraduate... Closing date: 18 Jan 2026 #compchem

"Enhancing Electrostatic Embedding for ML/MM Free Energy Calculations" is now out in #JCTC: pubs.acs.org/doi/10.1021/... Great job by João and team! #compchem

Enhancing Electrostatic Embedding for ML/MM Free Energy Calculations

Hybrid ML/MM approaches that combine machine learning (ML) potentials with molecular mechanics (MM) potentials offer a promising balance between computational cost and accuracy. Most ML/MM simulations reported to date employ mechanical embedding schemes, and rely on Lennard-Jones and Coulomb potentials to model intermolecular interactions between the ML and MM regions. A promising approach to improving ML/MM schemes is to use electrostatic embedding, where polarization effects on the ML region by the MM region are explicitly incorporated. The electrostatic machine learning embedding (EMLE) method has been developed for this purpose. Here, we compute absolute hydration free energies for a set of small organic molecules to derive robust methodologies for training EMLE models using quantum mechanical data. We establish protocols for fine-tuning the static and induced components of electrostatic interactions and evaluate the accuracy limits of fitting these components to first-principles calculations. We also introduce an empirical adjustment to enhance agreement with experimental results, strengthening the competitiveness of ML/MM simulations relative to state-of-the-art methods. Overall, our findings provide valuable insights into the challenges and opportunities of electrostatic embedding ML/MM simulations, and offer strategies for achieving robust modeling of classes of drug-like molecules where the accuracy of conventional MM force fields fall short.

pubs.acs.org

Cole Group@colegroupncl.bsky.social · last yr.

"Enhancing Electrostatic Embedding for ML/MM Free Energy Calculations", by Joao Morado et al, is now available on ChemRxiv! doi.org/10.26434/che... #compchem

Excited to be a part of this review paper by Chapin Cavender et al. Now out in LiveCoMS! doi.org/10.33011/liv...

If you're doing #alchemistry (alchemical free energy calculations) then here's release 2.5.0 of alchemlyb for you. github.com/alchemistry/... (There's also the alchemlyb @joss-openjournals.bsky.social paper doi.org/10.21105/jos... if you want to read & cite.)

2.5.0 · alchemistry alchemlyb · Discussion #445

New minor release of alchemlyb with fixes and enhancements. Supports Python 3.11 - 3.14. See CHANGES for details. What's Changed update action/checkout by @orbeckst in #417 Parallel read and prepro...

github.com

Preprint release 😀 of "Speak to a Protein," an AI co-scientist that facilitates data gathering and analysis in an interactive collaborative session. It is quite amazing to use. Preprint: arxiv.org/abs/2510.17826