Here's the last paper from my time in Frank Noe's group. We use normalizing flows (NF) to build a whole phase diagram from just two MD simulations. Especially useful for ML potentials, since it requires much fewer energy evaluations than traditional methods!

Neat. Should be much faster than creating it from scratch

blog: recreating an interesting visualization of electrostatic potentials with rdkit and matplotlib ljmartin.github.io/blog/25_esp_co…

🎉 Results of CACHE Challenge #2 are out! Five top-scoring teams from around the 🌍 used #AI, #CitizenScience, and #OpenSource tools to identify 7 promising hits that could lead to a pan-coronavirus treatment. Dataset is now publicly available to all! conscience.ca/news/cache-cha…

Very excited to have taken part in the CACHE2 hit finding challenge. Thank you so much to the organisers and well done to all the participants! We've written up some of our experiences in this blog: cole-group.github.io/2024/09/10/cac… Newcastle Chemistry Sciences at Newcastle

Pleased to see one of our compounds amongst the highest scoring in round 1, which we were able to follow up with free energy calculations in round 2. But more importantly we learned a lot about what we can improve when applying these methods in a prospective scenario

Switching between plotting libraries lets you leverage their unique strengths. However, learning multiple specific APIs can significantly increase your cognitive load. hvPlot provides a unified, high-level API that allows using the same syntax for Bokeh, Matplotlib, and Plotly.

There's been one paper I can't get out of my head for the past month, and it's Oliver Unke & Martin Stohr's work studying peptide and protein dynamics with NNPs. Here's why I think this work is so impactful, and what it means for the field: corinwagen.github.io/public/blog/20…

OpenDock: A PyTorch-Based Open-Source Framework for Protein-Ligand Docking and Modeling • A paradigm shift in molecular docking: OpenDock, built on PyTorch, offers an open-source framework with unparalleled flexibility for protein-ligand docking, integrating both classical and

We’ve just finished writing the missing 15,616 Wikipedia articles to get complete coverage of all 19,255 human genes. We used PaperQA2, which has higher accuracy than existing human-written Wikipedia articles, as judged by blinded biology PhD students and postdocs. 1/5

Introducing SRME: a new metric for further assessment of foundational machine learning interatomic potentials! 👉A smaller SRME value indicates better performance for calculating thermal conductivity 🖥️…tbench-discovery.materialsproject.org Huge thanks to Janosh Janosh & Yuan Yuan (Cyrus) Chiang

Eddie Millar Sam Dumitriu Hmm. We can put boundaries around it. NICE use ~£25k as value of extra quality year for humans. £100m is value of 4000 extra years. Bats live ~15 years, so that 266 bats if valued same as humans….. How big is the colony and how do you value bats vs humans?

Boltz-1: Democratizing Biomolecular Interaction Modeling 1. Boltz-1 introduces the first fully open-source model achieving AlphaFold3-level accuracy in predicting 3D structures of biomolecular complexes, democratizing access to cutting-edge structural biology tools. 2. The

this is lit. so satisfying to see conv nets in action man!

Today in nature, we’re presenting GenCast: our new AI weather model which gives us the probabilities of different weather conditions up to 15 days ahead with state-of-the-art accuracy. ☁️⚡ Here’s how the technology works. 🧵goo.gle/49trAOv

A new flagship quantum chemistry dataset is out! Omol25 contains more than 100 million DFT calculations across an expanse of chemical space. Collaborating with the FAIR team on this has been super fun! arxiv.org/abs/2505.08762

Transforming molecular machine learning with stereoelectronics-infused molecular graphs (SIMGs) Traditional molecular representations in machine learning often overlook essential quantum-mechanical effects, limiting predictive power in complex chemical systems. In my latest

You can estimate logarithms base 10 without a calculator with the approximation (x - 1)/(x + 1).

A surprising & little-known results in classical statistics: Mean (μ) and median (m) are within one std deviation: |μ−m| ≤ σ For unimodal densities, bound is even tighter |μ−m| ≤ 0.7746 σ This beautiful results first appeared in a 1932 paper by Hotelling & Solomons 1/3

📢 Call for proposals: Boltz small-molecule design collaboration! 🧬 Can we help design your ideal molecule? Can you help us improve our open-source models? Please reach out or share with scientists you know! More details below! It has been great to see the level of excitement