iiinova — where your digital self keeps evolving

Reposted by Pascal Notin

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 20/08/2026

Binder design is nice and all, but here we have three agents sharing data and autonomously controlling a lab to design enzymes with shifted substrate scopes and high activity! @cobanbrooks.bsky.social @pascalnotin.bsky.social @philromero.bsky.social www.biorxiv.org/content/10.6...

0205

Reposted by Pascal Notin

jproney @jproney.bsky.social · 13/03/2026

I'm excited to announce some major updates to our ProteinEBM paper with Chenxi Ou @sokrypton.org!

14111

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

Links: 🔗 Paper: www.biorxiv.org/content/10.1... 💻 Code: github.com/MarksLab-Das... 9/9

biorxiv.org

030

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

The moderate performance across all tasks reveals exciting opportunities! Key directions: RNA-specific training data, integrating structure-function relationships, and improving non-canonical base pair prediction. RNAGym provides the standardized foundation for progress. 7/9

100

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

🌀 Tertiary structure: 215 diverse 3D structures from the PDB. NuFold leads monomers (0.393 TM-score), AlphaFold3 dominates complexes (0.381 TM-score). Non-Watson-Crick interactions remain a major challenge for all methods 6/9

200

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

🔗 Secondary structure: 901k chemical mapping profiles using DMS & 2A3 reactivity. EternaFold achieves top performance (0.656 F1-score), closely followed by CONTRAfold & Vienna. Traditional thermodynamic methods are still competitive with newer deep learning approaches 5/9

120

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

🔬 Fitness prediction: 70 assays across tRNA, ribozymes, aptamers & mRNAs (1M+ mutations total). Evo 2 performs best overall (0.276), but performance varies dramatically by RNA type: RNA-FM excels at tRNA/aptamers while Evo 2 leads mRNA tasks. Lots of room for improvement across the board! 4/9

130

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

RNAGym tackles three essential RNA prediction tasks: 🔬 Fitness prediction: How mutations affect RNA function 🔗 Secondary structure: Base-pairing patterns 🌀 Tertiary structure: 3D molecular architecture All evaluated zero-shot to test true generalization! 3/9

110

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

Why do we need this? RNA modeling faces major challenges: limited experimental data (<1% of PDB entries), inherently less stable structures than proteins, and evaluation has been scattered across different studies with varying approaches. 2/9

110

Pascal Notin @pascalnotin.bsky.social · 18/06/2025

🚨 New paper 🚨 RNA modeling just got its own Gym! 🏋️ Introducing RNAGym, large-scale benchmarks for RNA fitness and structure prediction. 🧵 1/9

14016

Reposted by Pascal Notin

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 11/06/2025

End-to-end differentiable homology search for protein fitness prediction. @yaringal.bsky.social @deboramarks.bsky.social @pascalnotin.bsky.social arxiv.org/abs/2506.089...

0329

Pascal Notin @pascalnotin.bsky.social · 08/05/2025

But more broadly I wanted to convey in the blog that the two (structure + MSA) are critical for proper functional protein design & effects prediction

020

Pascal Notin @pascalnotin.bsky.social · 08/05/2025

Thank you @delalamo.xyz! Understand where you are coming from re: design. For some design setups structure is critical -- here my point was more for a directed evolution setup where you have to select top mutants that go in the next round

110

Pascal Notin @pascalnotin.bsky.social · 08/05/2025

Even simple methods leveraging these 2 modalities significantly outperform billion-parameter sequence-only models. So, what's next? Better retrieval, advanced multimodal approaches, & alignment. Read more: pascalnotin.substack.com/p/have-we-hi... #BioTech #AI #pLMs

pascalnotin.substack.com

Have We Hit the Scaling Wall for Protein Language Models?

Beyond Scaling: What Truly Works in Protein Fitness Prediction

181

Pascal Notin @pascalnotin.bsky.social · 08/05/2025

Have we hit a "scaling wall" for protein language models? 🤔 Our latest ProteinGym v1.3 release suggests that for zero-shot fitness prediction, simply making pLMs bigger isn't better beyond 1-4B parameters. The winning strategy? Combining MSAs & structure in multimodal models!

1247

Reposted by Pascal Notin

bioRxiv Biophysics @biorxiv-biophys.bsky.social · 25/03/2025

Large-scale discovery, analysis, and design of protein energy landscapes www.biorxiv.org/content/10.1101/202…

0128