Below the heads, the molecule becomes highly flexible and density is lost. However, using #XLMS and #AlphaFold, we were able to model these regions and identify bend 1 (A1156) as an unwinding point, enabling a full-length structure. (8/11)

I've repeatedly heard computer people and laypeople say AlphaFold solved protein folding. It's maddening.

Paper of the day · 16.04.26 Evaluating zero-shot prediction of monomeric protein design success by AlphaFold, ESMFold, and ProteinMPNN. De novo protein design has enabled the crea... Subnewsletter #ProteinDesign #StructuralBiology #Bioinformatics #ProteinEngineering

La base de datos que revolucionó la biología estructural acaba de dar el salto definitivo: ya no solo predice proteínas solitarias, sino cómo se unen para formar verdaderos “equipos” funcionales. Noticia en: www.ciencia1.com/2026/04/el-c... #AlphaFold #IAenBiologia #Ciencia #DeepMind #NVIDIA

Very nice account: @alphafold.bsky.social both recognises the success and undeniable utility of AlphaFold for what it is, and boosts callouts of it being overhyped and of it being used as the AI bro's go-to example for hyping up "AI" in general. Five stars for remaining grounded in reality.

go home everyone, AlphaFold solved the protein folding problem (essentially is doing a lot of work there) (this in a post from Semafor, which (appallingly) only links to a MIT press release about the relevant paper)

AlphaFold predates the current generation of AI. Effectiveness of cancer screenings still need to be evaluated. And obviously I’m not talking about ML in general here, which should be apparent based on my original post.

💡AlphaFold's pLDDT scores can be visualised in Jalview as annotation tracks at the bottom - useful when comparing regions of high confidence across predicted protein structures in alignments. ▶️ For a guided tour of AlphaFold metrics in Jalview, watch: www.youtube.com/watch?v=WVCM...

We’re delighted to share a ‘behind-the-scenes’ peek of SBS TV filming Matt Higgins @parasitematt.bsky.social and his group this week for a documentary on the Future of Medicine with AlphaFold. It will air in South Korea in early May 2026. Watch this space for more!

@mitsenkov.bsky.social , Senior Bioinformatician at AlphaFold DB, will be giving a keynote talk for EHA’s Computational Biology Training in Hematology programme, speaking with haematologists about structural biology and practical AlphaFold DB use-cases. #AlphaFold #EHA2026

Run an MD simulation of any protein in the AlphaFold Protein Structure Database using AF-CALVADOS Thanks to @sobuelow.bsky.social AF-CALVADOS is now on Colab colab.research.google.com/github/KULL-...

Great presentation from @ulad-litvin.bsky.social @microbiologysociety.org annual conference: How far can AlphaFold go? Lessons from modelling virus–host protein interactions using AlphaFold3.

Even the one good example of 'AI' in research is overhyped: AlphaFold does not model folding! It predicts final static structures. It tells us nothing about protein dynamics, and we haven't actually fully solved the folding problem until we can model protein dynamics.

Watch Live Now: NASA’s Artemis II Crew Comes Home (Official Broadcast) www.youtube.com/live/nfhDuOH... #science #artemis

ROCKET is giving the Nobel Prize-winning AI-powered AlphaFold a serious upgrade in its ability to predict how proteins fold: https://bit.ly/4dUTZkG #science #biology #FlatironCCM

🍿 "Our results suggest that, at this stage, the main contribution of AI predictions is to provide quaternary structure models for experimentally identified PPIs." #alphafold www.nature.com/articles/s41...

De la chiropratie pour protéines ou "C'est quoi AlphaFold ?" par @drpoloch.bsky.social youtube.com/shorts/vKxMd... par @bestofvulgatwitch.bsky.social #vulgarisation #sciences #biologie

@drpoloch.bsky.social nous explique le repliement des protéines et AlphaFold ! Retrouvez-la en stream sur Twitch : www.twitch.tv/polochsansgene

#FlatironCCM researchers have developed a new tool called ROCKET that extends the capabilities of AlphaFold by directly learning from raw experimental data: https://bit.ly/4dUTZkG #science #math

Worth sharing. Powerful. Please share. #AIForHumanity developer.nvidia.com/blog/designi...

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...

New paper from the Stern and Gronenborn labs on structure and evolution of aphid bicycle effector proteins. We show how alphafold can generate accurate predictions for these rapidly evolving proteins and that these proteins are diverse in every possible way. www.biorxiv.org/content/10.6...

AlphaFold as a prior: experimental structure determination conditioned on a pretrained neural network www.nature.com/articles/s41... rdcu.be/fbbiQ

A new software tool called ROCKET is giving AlphaFold a serious upgrade in its ability to predict how proteins fold: https://www.simonsfoundation.org/2026/04/01/rocket-upgrades-alphafold-to-learn-from-raw-experimental-data/

This series will guide you through accessing structural data programmatically with the Protein Data Bank in Europe (PDBe). The series also includes a final session on the AlphaFold Database (AFDB). Webinar series organisers: Ajay Mishra, Dr. Sudakshina Ganguly, and Flaminia Zane at @ebi.embl.org.

Starting on 6 May, our next webinar series will focus on different levels of programmatic access at @pdbeurope.bsky.social and the #AlphaFold DB. Registration for each webinar is free but essential: www.ebi.ac.uk/training/eve... 🧬🖥️📊

AI news of the day · 02.04.26 Introducing ROCKET: An Upgrade to AlphaFold That Learns From Raw Experimental Data - Simons Foundation Introducing ROCKET: An Upgrade to AlphaFold That Learns From Raw Experimental Da... Subnewsletter #ProteinDesign #StructuralBiology #Bioinformatics #ProteinEngineering

Excellent work @alisiafadini.bsky.social et al. - exciting stuff! #ROCKET #StructuralBiology #AlphaFold #ResearchSky #AcademicSky

We're a big step closer to automated determination of protein structures. The key? Having AlphaFold listen to experimental data. Great work, led by @alisiafadini.bsky.social and @minhuanli.bsky.social in an inspiring collaboration with @moalquraishi.bsky.social and @randyjread.bsky.social.

Starting from an #AlphaFold-Multimer prediction, we used #ROCKET to build a model of ZPD, a homopolymeric zona pellucida (#ZP) protein, into an initial #cryo-EM map at only ~9 Å resolution. A subsequently obtained 4.6 Å map highlighted how superior the ROCKET model was over the initial prediction:

This is getting traction on Bluesky, so sharing here in case it interests people: reminder that AlphaFold is a success of deep learning, but also of data sharing, curation, and data standards, without which it would have been impossible.

ROCKET Launches Structural Biology Past AlphaFold 🧪 🧬 #AcademicSky #higherEd #RNA #Protein www.mcb.harvard.edu/department/n... @hekstralab.bsky.social @cam.ac.uk @flatironinstitute.org @umeauniversitet.bsky.social@tcddublin.bsky.social @stjuderesearch.bsky.social @columbiauniversity.bsky.social

ROCKET makes AlphaFold context-aware. We iteratively steer prediction to agree with experiment (cryo-EM, crystallography) at inference time, no retraining. Structure determination becomes a search where ML priors and experiment productively combine and inform each other. 2/15

ROCKET 🚀 inference-time optimization of AlphaFold to fit structural data is published! rdcu.be/fa9YH Since our preprint, we’ve pushed it to regimes where other methods break: low resolution, weak signal, real experimental edge cases. Here’s what we learned: 1/15

This! AlphaFold is PDB + UniProt + Deep learning. No data sharing, no curation, no data standards, no AlphaFold.

AI only "works ✨" where there is rich data. Alphafold is a great example! the models that use the highly structured crystallography and evolutionary data (alphafold2) need WAY less data & do better then Protein language models that blindly gobble up all the unstructured protein data (ESMfold).

David is presenting a (Alphafold 2 driven) detective work identifying Ger protein domains involved in channel formation and mutant variants with constitutive germination See their Science paper www.science.org/doi/10.1126/... #KNVM2026

This. Would AlphaFold have happened if the PDB didn’t exist - say if you had to pay to license protein crystal structure databases, and derived data couldn’t be used without approval of the PDB. Not that I’m thinking of any other crystallography database…

My main gripe with the alphafold example is how it shows you need decades and decades of high quality data, well structured, open and accessible to train a model -- and yet they always gloss over it and pretend it's just AI and magic. No, we need to continuously invest in real data and FAIR data.

No, you can't trust it. Yes, you need to validate using other methods. Those other methods involve doing the work the long way with precision. Which means you don't need AlphaFold. So...it doesn't save us any time. Or rather, on some projects it WOULD! I just don't know which.

For me, the most maddening thing is when AI boosters, or detractors, all seem to agree that AlphaFold is an unmitigated good and useful tool showing the power of AI. I disagree. It's hit and miss. And the misses are REALLY expensive. Especially when the conversation suggests it is ALL HIT.

AlphaFold Database expands to proteome-scale quaternary structures www.biorxiv.org/cont... --- #proteomics #prot-preprint

New from @depaceraffa.bsky.social from our lab #NICHD #NIH with Carlos Dominguez-Gonzalez, Adeline Vanderver @childrensphila.bsky.social et al.: pathogenic BLOC1S1 variants disrupt BLOC-1/BORC complexes in children with severe neurodevelopmental disorder authors.elsevier.com/c/1mqGAgeXHm-E

Paper of the day · 25.03.26 Integrative structural and physicochemical characterization of chalcone synthase enzymes from medicinal plants using AlphaFold, molecular dock... Subnewsletter #ProteinDesign #StructuralBiology #Bioinformatics #ProteinEngineering

Guardians of large cancer genomics databases are restricting access to their data in very unhelpful ways. If we in the structural biology community had done that with our data there would never have been an AlphaFold or Rosetta.

Chemistry & Molecular Modeling MCP Servers (3.5/5) ChatMol (~89 stars) connects PyMOL and ChimeraX to AI agents for molecular visualization. RDKit gets two MCP servers for cheminformatics via natural language. 25+ servers: PubChem, ChEMBL, DrugBank, AutoDock Vina, AlphaFold, LAMMPS. Our review:

AlphaFold database now includes 1.7 million 'homodimers' - comprising two interacting strands of the same molecule go.nature.com/4sjP69b

Abgefahren!! #KIImpfstoff für krebskranken #Hund mit #ChatGPT und #AlphaFold #AI #mRMA #Vaccine news.unsw.edu.au/en/meet-the-...

This man used AI to create personalized cancer medicine for his dying dog - and it worked. Paul, a data engineer with zero biology background, used ChatGPT and AlphaFold to design a custom vaccine for his dog Rosie after vets gave her 6 months to live. #Startup #Entrepreneurship

I use a combination of Multalin and Espript. Then import the AlphaFold prediction of the protein of interest and it generates a fairly nice image with the sequence and AF-derived secondary structure overlaid.