Martin Pacesa
@martinpacesa.bsky.social
Assistant Professor at the University of Zurich.
🖥️ protein design, machine learning🤖, crystallography💎, cryoEM🔬. Avid weirdness connoisseur 🎩
🖥️ protein design, machine learning🤖, crystallography💎, cryoEM🔬. Avid weirdness connoisseur 🎩
Reposted by Martin Pacesa
In 2017 we showed that homomers evolve on the edge of supramolecular self-assembly. Here we show how to exploit this vulnerability so that ALL homomers can become bona fide drug-🎯. Delighted that Aim 3 of our @erc.europa.eu finally polymerized! Manuscript: 👉 rdcu.be/e27bT
February 9, 2026 at 5:36 PM
In 2017 we showed that homomers evolve on the edge of supramolecular self-assembly. Here we show how to exploit this vulnerability so that ALL homomers can become bona fide drug-🎯. Delighted that Aim 3 of our @erc.europa.eu finally polymerized! Manuscript: 👉 rdcu.be/e27bT
Reposted by Martin Pacesa
Introducing The Structural History of Eukarya (SHE): The first proteome-scale phylogeny constructed entirely from 3D structure.
We computed 300 trillion alignments across 1,542 species to map the tree of life. 🧵👇 (1/5)
We computed 300 trillion alignments across 1,542 species to map the tree of life. 🧵👇 (1/5)
February 7, 2026 at 8:50 AM
Introducing The Structural History of Eukarya (SHE): The first proteome-scale phylogeny constructed entirely from 3D structure.
We computed 300 trillion alignments across 1,542 species to map the tree of life. 🧵👇 (1/5)
We computed 300 trillion alignments across 1,542 species to map the tree of life. 🧵👇 (1/5)
Great analysis of our paper, thank you Engin!
On the lack of more biohysics: at that point it was pretty well established that these mini binders are super thermostable and highly expressed, so we chose not to include it.
On the lack of more biohysics: at that point it was pretty well established that these mini binders are super thermostable and highly expressed, so we chose not to include it.
February 3, 2026 at 9:05 AM
Great analysis of our paper, thank you Engin!
On the lack of more biohysics: at that point it was pretty well established that these mini binders are super thermostable and highly expressed, so we chose not to include it.
On the lack of more biohysics: at that point it was pretty well established that these mini binders are super thermostable and highly expressed, so we chose not to include it.
Really nice, I always hated the recursive argument in transformer papers that just because you now have lower perplexity that your model is better.
February 3, 2026 at 9:02 AM
Really nice, I always hated the recursive argument in transformer papers that just because you now have lower perplexity that your model is better.
Reposted by Martin Pacesa
Bluesky doesn’t really shine yet as a place to discover papers, especially at the intersection of biology and ML. To help a little, I’m going to start sharing papers I find interesting. To kick things off, here’s one from @petar-v.bsky.social and colleagues at DeepMind
arxiv.org/abs/2601.22950
arxiv.org/abs/2601.22950
Perplexity Cannot Always Tell Right from Wrong
Perplexity -- a function measuring a model's overall level of "surprise" when encountering a particular output -- has gained significant traction in recent years, both as a loss function and as a simp...
arxiv.org
February 3, 2026 at 8:47 AM
Bluesky doesn’t really shine yet as a place to discover papers, especially at the intersection of biology and ML. To help a little, I’m going to start sharing papers I find interesting. To kick things off, here’s one from @petar-v.bsky.social and colleagues at DeepMind
arxiv.org/abs/2601.22950
arxiv.org/abs/2601.22950
Even project grants, which were the lifeline of many labs, are now limited…does not bode well.
February 3, 2026 at 6:43 AM
Even project grants, which were the lifeline of many labs, are now limited…does not bode well.
Reposted by Martin Pacesa
🗓️ Save the Date!
The 5th European RosettaCon – Crossing Boundaries with Protein Design will take place in Lisbon, Portugal 🇵🇹
🗓️ October 28–30, 2026
Join the protein design community for an inspiring scientific meeting at the intersection of innovation and design.
The 5th European RosettaCon – Crossing Boundaries with Protein Design will take place in Lisbon, Portugal 🇵🇹
🗓️ October 28–30, 2026
Join the protein design community for an inspiring scientific meeting at the intersection of innovation and design.
February 3, 2026 at 6:05 AM
🗓️ Save the Date!
The 5th European RosettaCon – Crossing Boundaries with Protein Design will take place in Lisbon, Portugal 🇵🇹
🗓️ October 28–30, 2026
Join the protein design community for an inspiring scientific meeting at the intersection of innovation and design.
The 5th European RosettaCon – Crossing Boundaries with Protein Design will take place in Lisbon, Portugal 🇵🇹
🗓️ October 28–30, 2026
Join the protein design community for an inspiring scientific meeting at the intersection of innovation and design.
Here are the success rates of de novo pipelines based on which designs I could actually identify the methods for.
January 22, 2026 at 3:48 PM
Here are the success rates of de novo pipelines based on which designs I could actually identify the methods for.
Huge congratulations @nboyd.bsky.social with Mosaic that absolutely killed in the competition! 𝑩𝒊𝒏𝒅𝑪𝒓𝒂𝒇𝒕2 did also pretty well with the second highest hit rate in the competition!
proteinbase.com/collections/...
proteinbase.com/collections/...
proteinbase.com
January 22, 2026 at 8:29 AM
Huge congratulations @nboyd.bsky.social with Mosaic that absolutely killed in the competition! 𝑩𝒊𝒏𝒅𝑪𝒓𝒂𝒇𝒕2 did also pretty well with the second highest hit rate in the competition!
proteinbase.com/collections/...
proteinbase.com/collections/...
Reposted by Martin Pacesa
My time in @martinsteinegger.bsky.social's group is ending, but I’m staying in Korea to build a lab at Sungkyunkwan University School of Medicine. If you or someone you know is interested in molecular machine learning and open-source bioinformatics, please reach out. I am hiring!
mirdita.org
mirdita.org
Mirdita Lab - Laboratory for Computational Biology & Molecular Machine Learning
Mirdita Lab builds scalable bioinformatics methods.
mirdita.org
January 20, 2026 at 11:07 AM
My time in @martinsteinegger.bsky.social's group is ending, but I’m staying in Korea to build a lab at Sungkyunkwan University School of Medicine. If you or someone you know is interested in molecular machine learning and open-source bioinformatics, please reach out. I am hiring!
mirdita.org
mirdita.org
Reposted by Martin Pacesa
🔬 Postdoc position – Infection Biology & Cryo-ET (Zurich) 🇨🇭
We are looking for a motivated Postdoc to join our interdisciplinary team.
📩 Interested? Check out the job advert:
www.imm.uzh.ch/dam/jcr:cb17...
👉 More info on the lab: www.weiss-laboratory.com
Please share with interested colleagues!
We are looking for a motivated Postdoc to join our interdisciplinary team.
📩 Interested? Check out the job advert:
www.imm.uzh.ch/dam/jcr:cb17...
👉 More info on the lab: www.weiss-laboratory.com
Please share with interested colleagues!
January 19, 2026 at 1:15 PM
🔬 Postdoc position – Infection Biology & Cryo-ET (Zurich) 🇨🇭
We are looking for a motivated Postdoc to join our interdisciplinary team.
📩 Interested? Check out the job advert:
www.imm.uzh.ch/dam/jcr:cb17...
👉 More info on the lab: www.weiss-laboratory.com
Please share with interested colleagues!
We are looking for a motivated Postdoc to join our interdisciplinary team.
📩 Interested? Check out the job advert:
www.imm.uzh.ch/dam/jcr:cb17...
👉 More info on the lab: www.weiss-laboratory.com
Please share with interested colleagues!
Reposted by Martin Pacesa
Perfect birthday gift 🙌 happy this one is out!
A good end to 2025: rhodamine-binders for bioimaging in JACS: pubs.acs.org/doi/10.1021/...
Fast, Bright, and Reversible Fluorescent Labeling of Rhodamine-Binding Proteins
Rhodamine dyes conjugated to targeting ligands can yield exceptionally bright fluorescent probes for live-cell imaging. However, the limited permeability of such rhodamine derivatives restricts their ...
pubs.acs.org
December 31, 2025 at 2:03 PM
Perfect birthday gift 🙌 happy this one is out!
Fully circular double rainbow appeared today while snowboarding
December 30, 2025 at 2:54 PM
Fully circular double rainbow appeared today while snowboarding
Super excited about our 𝑩𝒊𝒏𝒅𝑪𝒓𝒂𝒇𝒕 pipeline being highlighted as one of @twistbioscience.com favourite publications of 2025!
We did synthesize all binder genes with Twist ❤️
We did synthesize all binder genes with Twist ❤️
BindCraft, an open-source and automated pipeline for de novo protein binder design is presented by Bruno Correia, Martin Pacesa, Lennart Nickel, and Christian Schellhaas:
“One-shot design of functional protein binders with BindCraft”
“One-shot design of functional protein binders with BindCraft”
One-shot design of functional protein binders with BindCraft - Nature
BindCraft, an open-source, automated pipeline for de novo protein binder design with experimental success rates of 10–100%, leverages AlphaFold2 weights to generate binders with nanomolar affinity wit...
www.nature.com
December 29, 2025 at 3:15 PM
Super excited about our 𝑩𝒊𝒏𝒅𝑪𝒓𝒂𝒇𝒕 pipeline being highlighted as one of @twistbioscience.com favourite publications of 2025!
We did synthesize all binder genes with Twist ❤️
We did synthesize all binder genes with Twist ❤️
Hail the chosen one
December 27, 2025 at 5:09 PM
Hail the chosen one
Repeat after me: GENE 👏 EXPRESSION 👏 PROFILING 👏 DOES 👏 NOT 👏 A 👏 VIRTUAL 👏 CELL 👏 MAKE 👏
December 26, 2025 at 12:11 PM
Repeat after me: GENE 👏 EXPRESSION 👏 PROFILING 👏 DOES 👏 NOT 👏 A 👏 VIRTUAL 👏 CELL 👏 MAKE 👏
Reposted by Martin Pacesa
I've been thinking about the "virtual cell" concept and wanted to write up a few thoughts. Specifically on how I think the prior experience in GWAS informs the most likely way these models will be useful.
andrewcarroll.github.io/2025/12/23/t...
andrewcarroll.github.io/2025/12/23/t...
The Virtual Cell Will Be More Like Gwas Than Alphafold
There has been significant discussion recently on the concept of the “virtual cell.” I want to summarize the key concepts regarding what the field wants from a virtual cell and the challenges we face....
andrewcarroll.github.io
December 24, 2025 at 1:52 AM
I've been thinking about the "virtual cell" concept and wanted to write up a few thoughts. Specifically on how I think the prior experience in GWAS informs the most likely way these models will be useful.
andrewcarroll.github.io/2025/12/23/t...
andrewcarroll.github.io/2025/12/23/t...
Reposted by Martin Pacesa
JOB ALERT: Operations Director (DCI–Lausanne).
The DCI, a joint initiative of EPFL, UNIL, UNIGE and UNIBE, is recruiting an Operations Director to lead the cryo-EM facility in Lausanne.
📩 Questions: CEO of the DCI: @chgenoud.bsky.social
🔗 Apply here: careers.epfl.ch/job/Lausanne...
The DCI, a joint initiative of EPFL, UNIL, UNIGE and UNIBE, is recruiting an Operations Director to lead the cryo-EM facility in Lausanne.
📩 Questions: CEO of the DCI: @chgenoud.bsky.social
🔗 Apply here: careers.epfl.ch/job/Lausanne...
December 17, 2025 at 12:00 PM
JOB ALERT: Operations Director (DCI–Lausanne).
The DCI, a joint initiative of EPFL, UNIL, UNIGE and UNIBE, is recruiting an Operations Director to lead the cryo-EM facility in Lausanne.
📩 Questions: CEO of the DCI: @chgenoud.bsky.social
🔗 Apply here: careers.epfl.ch/job/Lausanne...
The DCI, a joint initiative of EPFL, UNIL, UNIGE and UNIBE, is recruiting an Operations Director to lead the cryo-EM facility in Lausanne.
📩 Questions: CEO of the DCI: @chgenoud.bsky.social
🔗 Apply here: careers.epfl.ch/job/Lausanne...
Yeah, I dont think it’s a good think to show. Generating crap faster is not a bonus 😅
December 20, 2025 at 3:40 PM
Yeah, I dont think it’s a good think to show. Generating crap faster is not a bonus 😅
We could make a lot more trajectories pass if we wanted to, but we on purpose give it roadblocks along the way to make sure only very best interfaces pass. In a sense, every single hallucination iteration is a filtering step, while with diffusion you only have one step to get it right
December 20, 2025 at 8:45 AM
We could make a lot more trajectories pass if we wanted to, but we on purpose give it roadblocks along the way to make sure only very best interfaces pass. In a sense, every single hallucination iteration is a filtering step, while with diffusion you only have one step to get it right
The main difference are the self consistency steps along the way: 5 multimer models randomly swapped during design, interface must be robust enough to survive MPNN redesign, final AF2 monomer model filtering (most important), etc. Pyrosetta doesnt do much tbh
December 20, 2025 at 8:42 AM
The main difference are the self consistency steps along the way: 5 multimer models randomly swapped during design, interface must be robust enough to survive MPNN redesign, final AF2 monomer model filtering (most important), etc. Pyrosetta doesnt do much tbh
Reposted by Martin Pacesa
Are you interested in integrative structural biology, but feel a bit lost?
Don't worry, we have you covered with our FEBS Advanced Course, Lost in Integration Vol. 2 — probing biomolecules with AI and experiments
probingbiomolecules2026.febsevents.org
network.febs.org/posts/integr...
Don't worry, we have you covered with our FEBS Advanced Course, Lost in Integration Vol. 2 — probing biomolecules with AI and experiments
probingbiomolecules2026.febsevents.org
network.febs.org/posts/integr...
December 18, 2025 at 7:27 PM
Are you interested in integrative structural biology, but feel a bit lost?
Don't worry, we have you covered with our FEBS Advanced Course, Lost in Integration Vol. 2 — probing biomolecules with AI and experiments
probingbiomolecules2026.febsevents.org
network.febs.org/posts/integr...
Don't worry, we have you covered with our FEBS Advanced Course, Lost in Integration Vol. 2 — probing biomolecules with AI and experiments
probingbiomolecules2026.febsevents.org
network.febs.org/posts/integr...