1/ New preprint! with @solersanchisx.bsky.social @cwjpugh.bsky.social @federicobilleci.bsky.social @jonnyfrazer.bsky.social and Mafalda Dias, we introduce a scalable framework to improve stability prediction and separate folding from functional constraints. www.biorxiv.org/content/10.6...

Blending physics-based and inverse folding models to disentangle variant effects on stability and function

Protein sequences are constrained not only by the need to fold into stable structures, but also by specific functional requirements imposed by natural selection. Yet predictions of how amino-acid chan...

biorxiv.org

Mechanistic ODE models encode pathway topology but are blind to anything outside the pathway. Representation learning compresses data into low-dimensional embeddings but says nothing about mechanism. Deep Mechanistic Models (DMMs) are our attempt at both in one model www.biorxiv.org/content/10.6...

Deep Mechanistic Models reveal pathway-extrinsic drivers of mammary MAPK signalling heterogeneity

Cells sense and respond to their environment through signalling pathways, and the dynamics of these pathways shape cell fate even within genetically identical populations. Two largely separate computa...

biorxiv.org

So, bluesky, where is the discussion of the absolute insane things AI is doing in the realm of mathematics? It seems like we're nearly at the inflection point where the models are as good as human domain experts, and soon may well be better. Is that discourse only happening on twitter?

Using LLMs should allow us to think bigger, harder to grasp, difficult to execute, longer to develop, ideas. That idea that’s been in you brain since forever but you hadn’t had the time because you had to retool and what not. Trust yourself. Go for that shit. Now is the fucking time.

Clément Canonne@ccanonne.github.io · 4d ago

"We're all worried," as what it means to do research (in my field, Theoretical CS) seems to be shifting, and shifting fast. What to do? Senior researchers must lead by example, knowing that not everything will pan out. What I'm suggesting below may not work everywhere, but here's my own advice: 1/

I actually know the answer to this and it's slightly insane. Google pulls from Wikipedia to determine if a place is a city, town, neighbourhood, whatever. Brighton's first line describes it as a "seaside resort" so Google Maps thinks it's a neighbourhood and doesn't display it in the "city" font.

An important piece about how LLMs are exposing the rot in academic institutions, and we should take this seriously. A bit too confident in AI inevitability, but the focus on pre-existing issues is spot-on, from misaligned teaching to the rot of academic publishing. www.chronicle.com/article/the-...

The University After AI

The postwar bargain at the heart of higher education is coming undone. What replaces it?

chronicle.com

Online Now: Mechanism of MEK1 phosphorylation by the N-terminal acidic motif-mediated asymmetric BRAF dimer Online now:

Mechanism of MEK1 phosphorylation by the N-terminal acidic motif-mediated asymmetric BRAF dimer

Kondo, Notbohm, et al. show that the BRAF N-terminal acidic motif engages the dimeric interface to drive an asymmetric arrangement in which only one protomer adopts a canonical active kinase conformation. Captured MEK1 Ser222 phosphorylation and mutational analysis corroborate that this motif is required for full BRAF kinase activity.

dlvr.it

Review/perspective article by Diederik Laman Trip on some of his and our lab's thinking on how we can study human genetics across different levels. In particular around cell-type specific protein interactions, tissue organization and organ morphology. (paywall, shareable link rdcu.be/fvnpz)

Human genetics across levels of biological organization

Nature Reviews Genetics - Genetic variation influences biological processes across scales; however, deciphering how effects propagate from molecules to whole organisms remains a key challenge in...

rdcu.be

New lab preprint - Deep learning models are most often difficult to interpret black box models. What if we engineer them with interpretation in mind? Dennis Gankin carefully studied so-called biologically inspired neural networks and found an interesting phenomenon www.biorxiv.org/content/10.6...

Leveraging multiplicity in biologically informed neural networks to uncover disease heterogeneity

Biologically inspired neural networks (BINNs) embed pathway, ontology, or protein-interaction structure directly into neural networks, promising interpretable disease prediction where hidden nodes map...

biorxiv.org

The one thing I'll add (to those conclusions from people far more well informed than I: denial is not a solution (on either side). 4/n End

3) Best insight I've had from an AI expert: "Anyone who confidently asserts they know where this is all headed is wrong. No-one knows what the world will look like in a couple of years" 3/n

Our new paper is out in @nature.com! We uncovered how a naked mole-rat queen 👑 controls an entire colony: a natural contraceptive scent that suppresses reproduction in other females! A remarkable mammalian parallel to bees and ants! 🐝🐜 @garyrlewin.bsky.social youtu.be/enfv3zJrWKM?...

Naked mole-rats are ruled by a queen: this chemical puts her in power

YouTube video by nature video

youtu.be

Can we program a kinase like a switch? Inspired by natural autoinhibitory complexes, we designed miniproteins against active- and inactive-like conformations of Focal Adhesion Kinase. Depending on the targeted state, the resulting binders either activated or inhibited the kinase.

Our latest preprint is out. Great collaboration with the group of Jan Hasenauer (University of Bonn). A very comprehensive benchmarking of optimisation methods for parameter estimation in mechanistic models. Preprint in the post below.

bioRxiv SysBio@biorxiv-sysbio.bsky.social · 3w ago

Large-scale analysis of optimisation methods for parameter estimation problems in the life sciences https://www.biorxiv.org/content/10.64898/2026.07.11.737731v1

Please come and join us! Jobs for machine learning and mathy types!

UCL Research Department of Cell & Developmental Biology@ucl-cdb.bsky.social · 4w ago

📢 We're hiring! The Cell & Developmental Biology department (CDB) and other departments in Bioscience @ UCL are delighted to be recruiting for several Lecturer / Associate Professor positions. 👉 View full details and apply here: www.ucl.ac.uk/life-science...

The @qedscience.bsky.social "impact" score generated a lot of discussion on ranking preprints, including ideas on multi-dimensional rankings that are user specific. Besides describing the ideas, we can now prototype them (with Claude in this case). Here is Salient salient-sgeu6fs5ua-uc.a.run.app

Paying peer reviewers Interview with @danielgorelick.bsky.social about @biologyopen.bsky.social Fast & Fair trial in which peer reviewers were paid £220/review Spoiler: it works - fast, high quality reviews www.nature.com/articles/d41...

Why paying peer reviewers works, according to a journal’s editor-in-chief

A biology journal that paid peer reviewers found that the approach cut the time to a first editorial decision by 85% and maintained high-quality reviews.

nature.com

Claude Science - an AI workbench for science. This looks like an interesting attempt at making scientific workflows easier and more reproducible. I'm curious to give it a try. www.anthropic.com/news/claude-...

Claude Science, an AI workbench for scientists

Claude Science is a customizable app that integrates the tools and packages researchers most often use, produces auditable artifacts, and provides flexible access to computing resources.

anthropic.com

People always say "tell a story" but then don't explain how to do so! How is that helping? If we knew how to tell a story we wouldn't need the advice in the first place. 1/

Stephen Turner@stephenturner.us · last mo.

How to give a bad talk www.nature.com/articles/s41...

Our ambition to combine artificial intelligence with fundamental biology and medicine continues to evolve as we welcome Mihaela van der Schaar as our Chief AI Scientist. Read more: www.crick.ac.uk/news/2026-06...

Mihaela van der Schaar appointed Crick Chief AI Scientist

The Francis Crick Institute has appointed Professor Mihaela van der Schaar as its Chief AI Scientist, advancing its ambition to combine artificial intelligence with fundamental biology and medicine, c...

crick.ac.uk

Are cells worth having? Royal Society President Paul Nurse answers the hard questions youtu.be/7B4jko1VFP8?...

Philomena Cunk vs Sir Paul Nurse

YouTube video by Philomena Cunk

youtu.be

In our latest editorial @amjeve.uk & I take on the dreaded impact factor Like Monday mornings no one likes them but you can't ignore them & they keep coming around However there are things we should all do as authors, reviewers & editors journals.biologists.com/dev/article/...

The elephant in the room: impact factor

Summary: This Editorial offers an honest look at impact factor trends in developmental biology, explains how Development serves its community, and makes the case that sustaining developmental biology ...

journals.biologists.com

In 2022 Fantasy Premier League winner @joshuabull.bsky.social made his World Cup predictions. They were okay, but no cigar. However, in the intervening four years he has really honed his mathematical skills. Now he is 100% confident he's found the secret sauce.

One of the good things about the AI co-scientist brouhaha is that it makes us discuss what science *is*, and how to measure its value, impact, quality. — Questions that have been there all along, but many preferred to gloss over.

Introducing openRxiv Labs! 🔬 Growing from the strong foundation of bioRxiv & medRxiv, Labs is an experimental space for working with collaborators to push the boundaries of open scientific communication. 👉 Check it out and learn more: openrxiv.org/openrxiv-lab... #OpenScience #Preprints #Scicomm

Launching openRxiv Labs - openRxiv

Over the last thirteen years, bioRxiv and medRxiv have grown into widely used infrastructure for rapid research sharing in biology and medicine. The reliability and researcher-first values that have d...

openrxiv.org

"Hypothesis generation: pretty good...data analysis: a super power...executing experiments: not so much" AI can help at every stage of the research cycle. Jonah Cool from @anthropic.com discuss the benefits of integrating frontier models and potential concerns youtu.be/sz1vQyBzSgw?...

Interview with: Jonah Cool

YouTube video by CSHL Leading Strand

youtu.be

The people who believe that generative AI development & adoption everywhere is the ultimate goal & the first order predictor of success are insufferable. So are the people who think that AI can't ever produce anything useful and that everything they do produce is "slop". I miss reasoned arguments.

An amazing list of speakers and lots of opportunity to give selected talks! Save the date and join us in Amsterdam next year!

Stefano Di Talia@ditalialab.bsky.social · 3mo ago

Be on the lookout for an exciting meeting that Matilde @matildegalli.bsky.social , Silvia @santoslab.bsky.social and I are organizing next year in Amsterdam. It will be a lot of fun. coming-soon.embo.org/w27-24

Spurious correlation inflates performance in single-cell perturbation prediction www.biorxiv.org/content/10.6...

"a paper today is a Frankensteined-together patchwork of supplementary files that run longer than the manuscript, datasets in repositories...code posted to GitHub without documentation. The paper has been replaced....by a mess" www.thetransmitter.org/from-bench-t...

Is the scientific paper due to be replaced?

AI is pushing scientific publishing to the brink. For neuroscience, the crisis may be an opportunity to finally connect findings across subfields.

thetransmitter.org