iiinova — where your digital self keeps evolving

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

Different ranking and encoding, one of the stages has been changed as well to improve it, but otherwise pretty similar for basic hallucination.

010

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

No dependencies that need a license. Only restriction is if you are a compute service provider hosting BC2 code, otherwise free for industry for asset development and internal usage

010

Reposted by Martin Pacesa

Simon Bayly @simonbayly.bsky.social · 21/09/2026

ALT text “pink protein slides towards and then aggressively hugs smaller teal protein”

4155

Reposted by Martin Pacesa

Jase Gehring @skyjase.bsky.social · 21/09/2026

State of the art for open source binder design

063

Reposted by Martin Pacesa

Elizaveta Chernova @liza-ch.bsky.social · 21/09/2026

Check out BindCraft2! Working on this project was a great way to start my PhD.

0122

Reposted by Martin Pacesa

Dr. Märt-Erik Mäeots @erikmaeots.bsky.social · 21/09/2026

Great pleasure to have worked on this! Design peptides, miniproteins, megaproteins, nanobodies, ScFvs, anything you want all under one roof of BindCraft2. And it's easier, faster, and better than ever! Includes multitargeting against multiple different proteins as well.

1102

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

And ʙɪɴᴅᴄʀᴀꜰᴛ2 is for everyone: we've teamed up with @adaptyv.bio, AriaX Bio & Tamarind Bio to host it, so anyone can run it without a GPU. Built by @erikmaeots.bsky.social, @liza-ch.bsky.social , Leonardo Tredese & Jan Dernic. Special thanks to @sokrypton.org & Lennart Nickel. 🚀

1111

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

Want the full picture? It's all here: Overview + modalities: pacesalab.com/bindcraft Wiki + design guide: github.com/PacesaLab/Bi... Run it now on Colab: colab.research.google.com/github/Paces...

pacesalab.com

BindCraft2 | Pacesa Lab

BindCraft2 is the next generation of our platform for user-friendly de novo protein binder design. Like the original, it was built to make binder design easy…

1121

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

Already proven in competition: the top hit rate against the Nipah receptor-binding epitope, and the best of all approaches on RBX1. Much more wet-lab validation is coming in the preprint, with new features landing over the coming weeks. proteinbase.com/collections/...

proteinbase.com

GEM x Adaptyv: RBX1 Binder Design Competition - Results

Results from the GEM x Adaptyv RBX1 Binder Design Competition.

140

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

What we're most excited about: multi-target optimisation. One binder against several targets, with detargeting to avoid off-targets and paralogs. Lennart Nickel and Paul Kittner release a complementary cross-species approach that plugs onto BC1 for existing workflows. github.com/martinpacesa...

170

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

We built a custom integrated engine that makes ʙɪɴᴅᴄʀᴀꜰᴛ2 dramatically faster than v1. And it isn't limited to structured domains: you can target disordered regions, short linear motifs, or a target given as a bare sequence.

190

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

ʙɪɴᴅᴄʀᴀꜰᴛ2 now designs almost any binder format: • mini & large binders • linear & cyclic peptides • VHHs, scFvs & Fabs • ankyrin repeat proteins (ARPs) • homo-oligomers & multidomain binders • induced-fit & fold-switch binders

170

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

ʙɪɴᴅᴄʀᴀꜰᴛ2 is out, and we're not waiting for the paper. The full code drops today, free for academic and industry use. We're releasing it early so you can start designing right now, and bring its full power to the current Adaptyv competition. github.com/PacesaLab/Bi...

4292142

Martin Pacesa @martinpacesa.bsky.social · 21/09/2026

I also mix up Spanish and Japanese all the time, somehow in my brain they are equivalent.

010

Reposted by Martin Pacesa

Sergey Ovchinnikov @sokrypton.org · 14/09/2026

Finally a more intuitaive way to learn pLDDT/pAE? 😎 sokrypton.github.io/protein_figh... (Character idea from @hannes-stark.bsky.social & Alex Waldherr)

727698

Reposted by Martin Pacesa

Hauke Hillen @haukehillen.bsky.social · 11/09/2026

Happy to share our new preprint on the structure of the human mitochondrial RNA degradosome, a central player in organellar RNA metabolism! Work led by graduate student @paulafprado.bsky.social in our group. www.biorxiv.org/content/10.6... 🧵 1/9

19134

Reposted by Martin Pacesa

Janani Durairaj (Jay) @ninjani.bsky.social · 28/08/2026

Some exciting news: I'm joining the University of Lausanne @unil.bsky.social @dbc-unil.bsky.social as an Assistant Professor next month 🎉 My group will work on context-aware deep learning for protein structure, interaction & design. Postdoc and PhD openings coming soon - keep an eye out!

412232

Martin Pacesa @martinpacesa.bsky.social · 24/08/2026

I could do away with MD altogether 🤓

030

Martin Pacesa @martinpacesa.bsky.social · 22/08/2026

Thank you for your message. We are out of office until the 31st of August, with limited access to email and social media. We will get back to you at the earliest possible European working day.

050

Martin Pacesa @martinpacesa.bsky.social · 05/08/2026

I think their own wetlab validation will come in soon too

012

Martin Pacesa @martinpacesa.bsky.social · 04/08/2026

We are looking for a lab manager starting from January 2027, you can find more information on our members page pacesalab.com#members. We are also starting to look into hosting master students starting 2027.

11215

Reposted by Martin Pacesa

Martin Steinegger 🇺🇦 @martinsteinegger.bsky.social · 03/08/2026

Riboseek has now a WIP Marv Logo!

0335

Martin Pacesa @martinpacesa.bsky.social · 27/07/2026

Depenes what you consider msa-free. It language model based then even worse. Single sequence modes can be mutation sensitive though not always. Most structure predictors are not just due to training

011

Martin Pacesa @martinpacesa.bsky.social · 25/07/2026

Noooooooo, please not another one 😭😭😭😭 using MSAs for prediction will allow you to confidently fold any point mutant, it’s not actual validation

1160

Reposted by Martin Pacesa

Joe Greener @jgreener64.bsky.social · 17/07/2026

Function of a protein: something the protein does. Function in maths: a mapping from X to Y. Function in code: re-use code to do a task. Function at a conference: have a drink and look at posters.

092

Reposted by Martin Pacesa

Petr Skopintsev @petrskopintsev.bsky.social · 17/07/2026

Proud to present our work, published today in @science.org! My deepest gratitude to @isabelesain.bsky.social, @evandeturk.bsky.social, Jennifer and our collaborators in the Doudna, Cate, Banfield, and Jacobsen labs, and the @innovativegenomics.bsky.social! I hope you enjoy reading it! ✨

04310

Reposted by Martin Pacesa

Max Planck Institute for Medical Research @mpi-mr.bsky.social · 13/07/2026

📣 Join us for the next Rudolf Mößbauer Colloquium this Friday at 10 am at our institute! We are pleased to welcome @martinpacesa.bsky.social (University of Zurich)! Everyone interested is welcome to attend in person. #mpimr #maxplanck #RMC #Colloquium

Poster of Rudolf Mößbauer Colloquium with Martin Pacesa at MPI for Medical Research

091

Martin Pacesa @martinpacesa.bsky.social · 11/07/2026

The fold complexity scoring game has been updated with additional aspects of protein complexity according to suer feedback! All the scores have been recalculated and at least to me seem to correlate better with the apparent complexity of the fold! Happy voting!

072

Reposted by Martin Pacesa

Gregor Weiss @weiss-lab.bsky.social · 10/07/2026

Some bacteria do not live as individuals. They build multicellular filaments and connect neighboring cells through tiny cell-cell junctions. Cyanobacteria are among the most beautiful examples 😍 In our new paper, we reveal first molecuar insights on the cyanobacterial septal junction architecture 🧵

311527

Martin Pacesa @martinpacesa.bsky.social · 08/07/2026

Lots of votes already! Thank you everyone! I have again updated the set of structures based on the great feedback from people. You should have an easier time voting now and it includes more interesting structures.

041

Martin Pacesa @martinpacesa.bsky.social · 08/07/2026

One seems to be easier than the other

000

Martin Pacesa @martinpacesa.bsky.social · 08/07/2026

First update and score calibration, seems the base metric only agrees with humans 73.5% of the time! I have updated the set of structures which should make it more intuitive to vote. Hopefully we can improve the metric to a better agreement!

150

Reposted by Martin Pacesa

Rosetta Commons @rosettacommons.bsky.social · 07/07/2026

In the search to find a good metric for complex protein folds, the Pacesa Lab has created the FoldComplexity voting site. The results from weekly voting will be used to update subsamples of protein chains in the PDB to help better calibrate the metric. Check it out: pacesalab.com/foldcomplexity

052

Martin Pacesa @martinpacesa.bsky.social · 07/07/2026

There are currently several components based on human estimation of what makes a complex protein fold. Please send us feedback if you think we are missing something or currently severely underestimating one of the factors! Happy voting!

030

Martin Pacesa @martinpacesa.bsky.social · 07/07/2026

We have also added a subsample of 1000 protein chains form the PDB to help better calibrate the metric and see if one protein feels complex enough, relative to other folds. This will be updated weekly to account for your votes, until we reach a good value.

130

Martin Pacesa @martinpacesa.bsky.social · 07/07/2026

We often talk about complex protein folds but what does that really mean? Turns out we don't have a good metric for it. So we made a game to make one! Head over to our website to vote on which fold you think is more complex and why! pacesalab.com/foldcomplexi...

13417

Reposted by Martin Pacesa

Pedro Beltrao @pedrobeltrao.bsky.social · 06/07/2026

My department @eth-dbiol.bsky.social at ETH Zurich is hiring for a tenured professor in mol cell bio and biochemistry (associate/full). The position is within the Institute of Biochemistry (www.bc.biol.ethz.ch) and it is fairly open within their scope of research. www.nature.com/naturecareer...

nature.com

Professor of Molecular Cell Biology and Biochemistry - Zurich, Switzerland job with ETH Zurich | 12861386

The Department of Biology at ETH Zurich invites applications for the above-mentioned tenured professorship at the Institute of Biochemistry

07683

Reposted by Martin Pacesa

Khmelinskaia Lab @akhmelinlab.bsky.social · 05/07/2026

De novo designed oligomers that respond to copper, small molecules, and phosphorylation... Using a single design strategy? 🧬⚙️ 🎉 Excited to share our new bioRxiv preprint—a collaboration between the Khmelinskaia, Correia, Schoeder and a Tinnefeld labs! www.biorxiv.org/content/10.6...

biorxiv.org

A generalizable interface-seeded framework for de novo design of functional oligomers

Protein oligomers are ubiquitous in biological systems and essential for function. However, the de novo design of oligomers that controllably assemble in response to exogenous stimuli remains challeng...

2259

Reposted by Martin Pacesa

Brady Johnston @bradyajohnston.bsky.social · 03/07/2026

The paper is out which means the MD trajectories are finally available! #MolecularNodes #b3d #GeometryNodes

48919

Reposted by Martin Pacesa

Institute of Science and Technology Austria (ISTA) @istaresearch.bsky.social · 29/06/2026

AlphaFold—an AI tool that revolutionized structural biology and earned the 2024 Nobel Prize in Chemistry—was trained on static crystal structures that dominate protein datasets. However, proteins are highly dynamic molecules. Read more: ista.ac.at/en/news/towa...

ista.ac.at

Toward Experiment-Guided AlphaFold

The AI-based program AlphaFold predicts a protein’s 3D structure with remarkable accuracy. However, it tends to reduce heterogeneous structures to a single dominant conformation, or shape, and overloo...

1145

Reposted by Martin Pacesa

Institute of Science and Technology Austria (ISTA) @istaresearch.bsky.social · 29/06/2026

First author, ISTA PhD student Advaith Maddipatla. Co-leading authors: @alex-bronstein.bsky.social, Ailie Marx, @paulschanda.bsky.social, and @sankethvedula.bsky.social.

ISTA researchers guide AlphaFold with experimental data, paving the way for improved future predictive models. Left to right: Advaith Maddipatla, Meital Bojan, Alex Bronstein, Nadav Sellam Bojan, and Paul Schanda. © ISTA

174

Reposted by Martin Pacesa

Mikko Taipale @miketilapia.bsky.social · 30/06/2026

Our evil ORFeome paper is out! In collaboration with @alex-stark.bsky.social, we screened ~4,000 viral proteins and secreted effectors from bacteria and parasites, covering hundreds of diverse pathogens, for phenotypes in human cells.

29943

Reposted by Martin Pacesa

Chi-Min Ho @cmholab.bsky.social · 30/06/2026

Incredibly proud of Messi Haile @mehsehret.bsky.social, a founding member of the Ho Lab, for her beautiful work uncovering the inner workings of a central piece of the malaria parasite invasion machinery: the malarial moving junction, out today in @cellpress.bsky.social! tinyurl.com/4p3md2eb 🦠❄️🔬

410935

Martin Pacesa @martinpacesa.bsky.social · 30/06/2026

I have reviewed so many of these QTY papers and even wrote an extensive rebuttal on it in our own solubilisation paper (static-content.springer.com/esm/art%3A10...). Making mutations and predicting WITH MSAs is NOT validation, you can predict any nonsense sequence with low RMSD with MSA enabled.

static-content.springer.com

040

Martin Pacesa @martinpacesa.bsky.social · 24/06/2026

Bad timing to start my lab around the GTA VI release. At least Elder Scrolls 6 and Fallout 5 won’t come out before tenure evaluation.

1160

Reposted by Martin Pacesa

Oli Clarke @olibclarke.bsky.social · 24/06/2026

Amazing stuff, what a structure! Elegant work😍

2193