🚀New @IHIEurope project launched: LIGAND-AI.
Led by @thesgconline & @Pfizer_UK, the project will generate open protein-ligand interaction datasets to power AI models for drug discovery.
EMBL-EBI will archive data & integrate results into @ChEMBL.
www.ebi.ac.uk/about/news/t...
#OpenData
Led by @thesgconline & @Pfizer_UK, the project will generate open protein-ligand interaction datasets to power AI models for drug discovery.
EMBL-EBI will archive data & integrate results into @ChEMBL.
www.ebi.ac.uk/about/news/t...
#OpenData
Powering drug discovery through AI-ready protein–ligand data
LIGAND-AI set to generate open, standardised, high-quality datasets of protein–ligand interactions at an unprecedented scale.
www.ebi.ac.uk
January 26, 2026 at 4:29 PM
🚀New @IHIEurope project launched: LIGAND-AI.
Led by @thesgconline & @Pfizer_UK, the project will generate open protein-ligand interaction datasets to power AI models for drug discovery.
EMBL-EBI will archive data & integrate results into @ChEMBL.
www.ebi.ac.uk/about/news/t...
#OpenData
Led by @thesgconline & @Pfizer_UK, the project will generate open protein-ligand interaction datasets to power AI models for drug discovery.
EMBL-EBI will archive data & integrate results into @ChEMBL.
www.ebi.ac.uk/about/news/t...
#OpenData
If you are interested in postdoctoral research on ligand-gated ion channels and the synapse, or simulation/AI methods for structural biology - drop me a line if you'll be around at Biophysical Society in LA! #bps2025 #bps #biophysics
February 14, 2025 at 5:50 AM
If you are interested in postdoctoral research on ligand-gated ion channels and the synapse, or simulation/AI methods for structural biology - drop me a line if you'll be around at Biophysical Society in LA! #bps2025 #bps #biophysics
Is #AI hitting a plateau in structure prediction? Help us find out at CASP17! 🧪🧬
Calling for Targets: Immune Complexes, protein - ligand complexes, RNA/DNA, conformational ensembles, membrane proteins, viral origins, and large complexes.
The Rule of Thumb: If AF3 can’t model it, we want it.
Calling for Targets: Immune Complexes, protein - ligand complexes, RNA/DNA, conformational ensembles, membrane proteins, viral origins, and large complexes.
The Rule of Thumb: If AF3 can’t model it, we want it.
March 13, 2026 at 11:51 AM
Is #AI hitting a plateau in structure prediction? Help us find out at CASP17! 🧪🧬
Calling for Targets: Immune Complexes, protein - ligand complexes, RNA/DNA, conformational ensembles, membrane proteins, viral origins, and large complexes.
The Rule of Thumb: If AF3 can’t model it, we want it.
Calling for Targets: Immune Complexes, protein - ligand complexes, RNA/DNA, conformational ensembles, membrane proteins, viral origins, and large complexes.
The Rule of Thumb: If AF3 can’t model it, we want it.
Have #AI methods for protein-ligand co-folding moved beyond memorisation and predict really novel leads for #drugdiscovery? Or do we find “more of the same”? This new benchmark lets you find out… ⬇️⬇️⬇️
Excited to share our latest preprint evaluating AlphaFold3, Boltz-1, Chai-1 and Protenix for predicting protein-ligand interactions, featuring our newly introduced benchmark dataset 🌹Runs N’ Poses🌹!
www.biorxiv.org/content/10.1...
🧵👇 (1/n)
www.biorxiv.org/content/10.1...
🧵👇 (1/n)
Have protein-ligand co-folding methods moved beyond memorisation?
Deep learning has driven major breakthroughs in protein structure prediction, however the next critical advance is accurately predicting how proteins interact with other molecules, especially small mo...
www.biorxiv.org
February 8, 2025 at 10:53 AM
Have #AI methods for protein-ligand co-folding moved beyond memorisation and predict really novel leads for #drugdiscovery? Or do we find “more of the same”? This new benchmark lets you find out… ⬇️⬇️⬇️
Deducing generalizable principles is much harder, and I am not aware of evidence that AI excels at finding principles.
Have protein-ligand co-folding methods moved beyond memorisation?
www.biorxiv.org/content/10.1...
Have protein-ligand co-folding methods moved beyond memorisation?
www.biorxiv.org/content/10.1...
Have protein-ligand co-folding methods moved beyond memorisation?
www.biorxiv.org
April 18, 2025 at 12:15 PM
Deducing generalizable principles is much harder, and I am not aware of evidence that AI excels at finding principles.
Have protein-ligand co-folding methods moved beyond memorisation?
www.biorxiv.org/content/10.1...
Have protein-ligand co-folding methods moved beyond memorisation?
www.biorxiv.org/content/10.1...
Professor Michael Bronstein @mmbronstein.bsky.social and researchers from EPFL have used AI to predict ‘neosurfaces’ that form when drugs bind to proteins. This breakthrough could accelerate precision medicine by using molecular glues to influence protein behaviour. www.nature.com/articles/s41...
Targeting protein–ligand neosurfaces with a generalizable deep learning tool - Nature
A computational deep learning approach is used to design synthetic proteins that target the neosurfaces formed by protein–ligand interactions, with applications in the development of new therapeutic m...
www.nature.com
January 17, 2025 at 3:24 PM
Professor Michael Bronstein @mmbronstein.bsky.social and researchers from EPFL have used AI to predict ‘neosurfaces’ that form when drugs bind to proteins. This breakthrough could accelerate precision medicine by using molecular glues to influence protein behaviour. www.nature.com/articles/s41...
Much like the Island Domain shifting its shape, our preprint has transformed too @natcomms.nature.com rdcu.be/eTOhF
β-sheet stabilization of the island domain underlies ligand-induced LRR-RP activation of plant immune signaling
Nature Communications - Using AI-based structural modeling, the authors reveal the ligand-binding and subsequent receptor complex formation mechanisms of three LRR-RP-type plant PRRs and propose a...
rdcu.be
December 9, 2025 at 9:01 AM
Much like the Island Domain shifting its shape, our preprint has transformed too @natcomms.nature.com rdcu.be/eTOhF
Excited to see more biology open-source models for real positive use-cases of AI!
Chai does structure predictions at AlphaFold3 levels of accuracy and able to handle multi-peptide or peptide-ligand complexes rather than just single chains.
Apache 2.0 on HF huggingface.co/chaidiscover...
Chai does structure predictions at AlphaFold3 levels of accuracy and able to handle multi-peptide or peptide-ligand complexes rather than just single chains.
Apache 2.0 on HF huggingface.co/chaidiscover...
December 5, 2024 at 2:39 PM
Excited to see more biology open-source models for real positive use-cases of AI!
Chai does structure predictions at AlphaFold3 levels of accuracy and able to handle multi-peptide or peptide-ligand complexes rather than just single chains.
Apache 2.0 on HF huggingface.co/chaidiscover...
Chai does structure predictions at AlphaFold3 levels of accuracy and able to handle multi-peptide or peptide-ligand complexes rather than just single chains.
Apache 2.0 on HF huggingface.co/chaidiscover...
🧪 AI programs like AlphaFold and RoseTTAFold can predict 3D protein structures but struggle to provide robust evidence about ligand binding crucial for drug discovery. Work in modeling protein structures bound to a ligand continues, but lab validation remains crucial. 🧬🖥️
AlphaFold touted as next big thing for drug discovery — but is it?
Questions remain about whether the AI tool for predicting protein structures can really shake up the pharmaceutical industry. Questions remain about whether the AI tool for predicting protein structur...
www.nature.com
September 25, 2023 at 4:39 PM
🧪 AI programs like AlphaFold and RoseTTAFold can predict 3D protein structures but struggle to provide robust evidence about ligand binding crucial for drug discovery. Work in modeling protein structures bound to a ligand continues, but lab validation remains crucial. 🧬🖥️
The link to the open one is here, if anyone reading this is interested ;-)
www.wur.nl/en/vacancy/p...
www.wur.nl/en/vacancy/p...
PhD student position in AI for life sciences with a focus on protein-ligand interaction dynamics
www.wur.nl
March 26, 2025 at 9:54 PM
The link to the open one is here, if anyone reading this is interested ;-)
www.wur.nl/en/vacancy/p...
www.wur.nl/en/vacancy/p...
New post about a very interesting new paper: When it comes to predicting protein-ligand binding, AI don't trump physics. medchemash.substack.com/p/ai-dont-tr...
AI don't trump physics
First principles remain important
medchemash.substack.com
October 8, 2025 at 7:57 PM
New post about a very interesting new paper: When it comes to predicting protein-ligand binding, AI don't trump physics. medchemash.substack.com/p/ai-dont-tr...
AI-enhanced adaptive virtual screening of large libraries for ligand discovery - @cgorgulla.bsky.social www.nature.com/articles/s41...
AI-enhanced adaptive virtual screening of large libraries for ligand discovery - Nature Biotechnology
Massive chemical libraries are screened computationally at scale.
www.nature.com
September 1, 2026 at 2:30 PM
AI-enhanced adaptive virtual screening of large libraries for ligand discovery - @cgorgulla.bsky.social www.nature.com/articles/s41...
The presence of buried waters in a binding pocket is a major determinant of how well Boltz-2 can predict binding affinities
www.linkedin.com/pulse/explor...
www.linkedin.com/pulse/explor...
Exploring Boltz-2’s Performance with the Uni-FEP Benchmark Dataset
The Atombeat team is excited to see the recent release of Boltz-2[1], an AI-driven algorithm for protein-ligand binding structure prediction and binding affinity estimation. As a computational drug di...
www.linkedin.com
June 29, 2025 at 11:33 AM
The presence of buried waters in a binding pocket is a major determinant of how well Boltz-2 can predict binding affinities
www.linkedin.com/pulse/explor...
www.linkedin.com/pulse/explor...
# AI in Drug discovery just BROKE THROUGH a wall #
A newer AI model, ConGLUDe, as fast but much more accurate than DrugCLIP.
Instead on just 40K structure-based data, ConGLUDe is trained on 100M datapoints from ligand-based data
P: arxiv.org/abs/2601.09693
A newer AI model, ConGLUDe, as fast but much more accurate than DrugCLIP.
Instead on just 40K structure-based data, ConGLUDe is trained on 100M datapoints from ligand-based data
P: arxiv.org/abs/2601.09693
January 15, 2026 at 1:13 PM
# AI in Drug discovery just BROKE THROUGH a wall #
A newer AI model, ConGLUDe, as fast but much more accurate than DrugCLIP.
Instead on just 40K structure-based data, ConGLUDe is trained on 100M datapoints from ligand-based data
P: arxiv.org/abs/2601.09693
A newer AI model, ConGLUDe, as fast but much more accurate than DrugCLIP.
Instead on just 40K structure-based data, ConGLUDe is trained on 100M datapoints from ligand-based data
P: arxiv.org/abs/2601.09693
🧱 New models like AlphaFold 3, Boltz-1, and PocketGen are transforming protein-ligand interactions, while RhoFold predicts 3D RNA structures.
💉 AI agents are even autonomously designing nanobodies against SARS-CoV-2, paving the way for AI-led discovery.
3/4
💉 AI agents are even autonomously designing nanobodies against SARS-CoV-2, paving the way for AI-led discovery.
3/4
February 3, 2025 at 6:36 AM
🧱 New models like AlphaFold 3, Boltz-1, and PocketGen are transforming protein-ligand interactions, while RhoFold predicts 3D RNA structures.
💉 AI agents are even autonomously designing nanobodies against SARS-CoV-2, paving the way for AI-led discovery.
3/4
💉 AI agents are even autonomously designing nanobodies against SARS-CoV-2, paving the way for AI-led discovery.
3/4
More protein-ligand data are needed for AlphaFold-like models (& AI/ML) to enable prospective design!
Check out our piece in "Current Opinion in Structural Biology" – Equal parts a thank-you-letter to the PDB & summarizes how datasets will enable task-focused models!
Link: doi.org/10.1016/j.sb...
Check out our piece in "Current Opinion in Structural Biology" – Equal parts a thank-you-letter to the PDB & summarizes how datasets will enable task-focused models!
Link: doi.org/10.1016/j.sb...
April 2, 2026 at 7:41 PM
More protein-ligand data are needed for AlphaFold-like models (& AI/ML) to enable prospective design!
Check out our piece in "Current Opinion in Structural Biology" – Equal parts a thank-you-letter to the PDB & summarizes how datasets will enable task-focused models!
Link: doi.org/10.1016/j.sb...
Check out our piece in "Current Opinion in Structural Biology" – Equal parts a thank-you-letter to the PDB & summarizes how datasets will enable task-focused models!
Link: doi.org/10.1016/j.sb...
Though the area has seen rapid growth, there are still caveats with AI-based docking methods. In this #ChemSci, Charlotte Deane and colleagues become “PoseBusters”, finding that AI docking methods struggle to generalise novel sequences and to generate physically valid poses.
PoseBusters: AI-based docking methods fail to generate physically valid poses or generalise to novel sequences
The last few years have seen the development of numerous deep learning-based protein–ligand docking methods. They offer huge promise in terms of speed and accuracy. However, despite claims of…
doi.org
March 10, 2025 at 3:33 PM
Though the area has seen rapid growth, there are still caveats with AI-based docking methods. In this #ChemSci, Charlotte Deane and colleagues become “PoseBusters”, finding that AI docking methods struggle to generalise novel sequences and to generate physically valid poses.
LIGAND-AI led by @thesgc.bsky.social at @uhnresearch.ca and funded by the Innovative Health Initiative (IHI) with EU-Canada support, is officially launched!
My team will contribute AI expertise and will develop the AIRCHECK platform (aircheck.ai) for drug discovery.
www.thesgc.org/news/ihi-pro...
My team will contribute AI expertise and will develop the AIRCHECK platform (aircheck.ai) for drug discovery.
www.thesgc.org/news/ihi-pro...
January 20, 2026 at 2:52 PM
LIGAND-AI led by @thesgc.bsky.social at @uhnresearch.ca and funded by the Innovative Health Initiative (IHI) with EU-Canada support, is officially launched!
My team will contribute AI expertise and will develop the AIRCHECK platform (aircheck.ai) for drug discovery.
www.thesgc.org/news/ihi-pro...
My team will contribute AI expertise and will develop the AIRCHECK platform (aircheck.ai) for drug discovery.
www.thesgc.org/news/ihi-pro...
Database for extended ligand-target analyses (DELTA): a new balanced resource for AI applications in drug discovery [new]
Balanced data enabled ligand/target analyses, showing activity trends.
Balanced data enabled ligand/target analyses, showing activity trends.
May 27, 2025 at 6:05 PM
Database for extended ligand-target analyses (DELTA): a new balanced resource for AI applications in drug discovery [new]
Balanced data enabled ligand/target analyses, showing activity trends.
Balanced data enabled ligand/target analyses, showing activity trends.
"Improvements in protein structure prediction are widely assumed to lead to better models of ligand-binding pockets and thus more accurate predictions of ligand binding (Beuming and Sherman, 2012; Bordogna et al., 2011; Erickson et al., 2004), but to what extent do these assumptions hold?"
How accurately can one predict drug binding modes using AlphaFold models?
AlphaFold 2 models capture binding pocket structures accurately but fare poorly when used to predict ligand-binding poses.
elifesciences.org
December 25, 2023 at 1:56 PM
CASP17, the 17th Critical Assessment of Structure Prediction conference, will take place in person in Rome, Italy, December 1–5, 2026. Registration is open now.
What is the state-of-the-art in predicting protein and RNA structures, immune complexes, and ligands?
events.humanitix.com/casp17 🧪 #AI
What is the state-of-the-art in predicting protein and RNA structures, immune complexes, and ligands?
events.humanitix.com/casp17 🧪 #AI
CASP17
CASP17 - Critical Assessment of Structure Prediction
3D Structure of Proteins, RNA, and Complexes, Immune Complexes, Ligand Binding, Conformational Ensembles
events.humanitix.com
July 11, 2026 at 8:13 AM
CASP17, the 17th Critical Assessment of Structure Prediction conference, will take place in person in Rome, Italy, December 1–5, 2026. Registration is open now.
What is the state-of-the-art in predicting protein and RNA structures, immune complexes, and ligands?
events.humanitix.com/casp17 🧪 #AI
What is the state-of-the-art in predicting protein and RNA structures, immune complexes, and ligands?
events.humanitix.com/casp17 🧪 #AI
PhD position in AI focusing on protein-ligand interaction dynamics #ScienceJobs #AcademicSky jobrxiv.org/job/wage...
PhD position in AI focusing on protein-ligand interaction dynamics
Post a job in 3min, or find thousands of job offers like this one at jobRxiv!
jobrxiv.org
March 30, 2025 at 3:12 PM
PhD position in AI focusing on protein-ligand interaction dynamics #ScienceJobs #AcademicSky jobrxiv.org/job/wage...
The future of electronics isn’t one device or material.
Build next-gen #sensors, memory & #photonic devices with NANOARC ligand-free #quantum material nanoinks. #Silicene #nanotubes, NbOx, #Gold & #2Dmaterials optimised for 3D & #PrintedElectronics
quantum.nanoarc.org/services/nan...
#DeepTech #AI
Build next-gen #sensors, memory & #photonic devices with NANOARC ligand-free #quantum material nanoinks. #Silicene #nanotubes, NbOx, #Gold & #2Dmaterials optimised for 3D & #PrintedElectronics
quantum.nanoarc.org/services/nan...
#DeepTech #AI
June 4, 2026 at 8:59 PM
The future of electronics isn’t one device or material.
Build next-gen #sensors, memory & #photonic devices with NANOARC ligand-free #quantum material nanoinks. #Silicene #nanotubes, NbOx, #Gold & #2Dmaterials optimised for 3D & #PrintedElectronics
quantum.nanoarc.org/services/nan...
#DeepTech #AI
Build next-gen #sensors, memory & #photonic devices with NANOARC ligand-free #quantum material nanoinks. #Silicene #nanotubes, NbOx, #Gold & #2Dmaterials optimised for 3D & #PrintedElectronics
quantum.nanoarc.org/services/nan...
#DeepTech #AI
Except the AI/ML guys in computational protein and ligand design who aren't themselves subject to the hate that breeds on X. They are mostly over there and not here, which is sad.
Some of them are on linkedin which now has some long threads on preprints/papers and some scientific debates.
Some of them are on linkedin which now has some long threads on preprints/papers and some scientific debates.
Bluesky is the new science Twitter, new study by @whysharksmatter.bsky.social and Julia Wester concludes!
"Results show that for every reported professional benefit that scientists once gained from Twitter, scientists can now gain that benefit more effectively on Bluesky than on Twitter."
"Results show that for every reported professional benefit that scientists once gained from Twitter, scientists can now gain that benefit more effectively on Bluesky than on Twitter."
Scientists no Longer Find Twitter Professionally Useful, and have Switched to Bluesky
Synopsis. Social media has become widely used by the scientific community for a variety of professional uses, including networking and public outreach. For
academic.oup.com
February 14, 2026 at 6:05 AM
Except the AI/ML guys in computational protein and ligand design who aren't themselves subject to the hate that breeds on X. They are mostly over there and not here, which is sad.
Some of them are on linkedin which now has some long threads on preprints/papers and some scientific debates.
Some of them are on linkedin which now has some long threads on preprints/papers and some scientific debates.
Run Openfold3 on my laptop machine #openfold3 #co-folding #informatics
Recently several co-folding AI are available, for example AF3, Chai-1, protenix and openfold. This area is hot and growing day by day. Co-Folding AI has potential for taking SBDD approach without determining the target…
Recently several co-folding AI are available, for example AF3, Chai-1, protenix and openfold. This area is hot and growing day by day. Co-Folding AI has potential for taking SBDD approach without determining the target…
Run Openfold3 on my laptop machine #openfold3 #co-folding #informatics
Recently several co-folding AI are available, for example AF3, Chai-1, protenix and openfold. This area is hot and growing day by day. Co-Folding AI has potential for taking SBDD approach without determining the target protein-ligand structure. As most of readers know that there are still limitation of using co-folding AI for SBDD but really useful I think. I have an interest in openfold because openfold disclose code and has good open scientific community.
iwatobipen.wordpress.com
August 1, 2026 at 5:38 AM
Run Openfold3 on my laptop machine #openfold3 #co-folding #informatics
Recently several co-folding AI are available, for example AF3, Chai-1, protenix and openfold. This area is hot and growing day by day. Co-Folding AI has potential for taking SBDD approach without determining the target…
Recently several co-folding AI are available, for example AF3, Chai-1, protenix and openfold. This area is hot and growing day by day. Co-Folding AI has potential for taking SBDD approach without determining the target…