After yet another tour through a whole stack of Python workflow systems, I still can't find one that beats Nextflow for #bioinformatics. Here's a short thread on the fatal flaws of each:

We (@XiaoWan38018817 and I) recently organized a special issue on Machine Learning in Chemical Biology in Current Opinion in Chemical Biology -- see our editorial for an overview of the eight fantastic contributions: authors.elsevier.com/a/1dr-u4sz6MBA…

Deep learning has obviously transformed protein structure prediction, but can it do the same for the rest of biology? In a perspective by Peter Sorger and myself out this week in Nature Methods, we begin to try to answer this question: rdcu.be/cyMvO

This is by far the best perspective piece that exposes very clearly what the deep learning toolkit can offer to modeling biology. If u think deep learning is just mindlessly fitting random MLP/CNN/LSTMs to data, you REALLY need to read this. 1/

It's finally here: the DeepTime software library!!! Massive work led by Moritz Hoffmann, with Martin Scherer, Tim Hempel, Andreas Mardt, Brian de Silva, BH, Stefan Klus, Hao Wu, Nathan Kutz and Steven Brunton. arxiv.org/abs/2110.15013

Jointly with Arne Elofsson @arneelof.bsky.social's group we applied Alphafold to predict complex structures for 65,000 human protein-protein interactions. Here is what we learned in trying to generate a structurally resolved human interaction network biorxiv.org/content/10.110…

An announcement I’ve been aching to make! After much sweat, we’ve built a trainable version of AlphaFold2, implemented in PyTorch, which we’re calling OpenFold. GitHub: github.com/aqlaboratory/o… Colab: colab.research.google.com/github/aqlabor… Why a trainable version of AlphaFold2 you ask? ⬇️

Took a look at the spike mutations in B.1.1.529 this evening, and colour coded them (details below)...there is...not much green.🧵

Nice demonstration of integrating genetic annotations with alphafold predicted structures: community.opentargets.org/t/how-to-overl…

On Thanksgiving, I am grateful for my family, delicious food, and Heng Li's foundation-laying code contributions to genomics. Cheers to all the coding scientists, whose shoulders we stand on everyday!

For most types of deep learning papers, it has become entirely viable to eschew conferences and journals and instead stick to arXiv only. That's because there's an objective grounding function that provides a much better feedback signal than publication ever could: adoption.

An annual round of predictions in Geometric and Graph ML coauthored with Petar Veličković based on input from the leading experts in the field. Longread on Towards Data Science towardsdatascience.com/predictions-an…

Our article “How does a small molecule bind at a cryptic binding site?” has been published in PLOS Computational Biology. doi.org/10.1371/journa…

It's incredible how quickly machine learning advances are democratized to a broader audience. Great to see collaboration between Richard Evans, Tim Green (AlphaFold) and @thesteinegger (ColabFold).

The Machine Learning in Structural Biology workshop MLSB (in San Diego) is on again NeurIPS Conference, in-person/hybrid in New Orleans this Dec! Stay tuned for our call for papers 📜 More info here: mlsb.io Sign up for our mailing list: groups.google.com/g/workshopmlsb

Forbes made me change the title (and some of the text) because it was too critical. They are doing this regularly of late, although my column is clearly labelled as opinion. New title: "Moderna Sues Rivals Over Covid-19 Vaccine Patent Rights, After Saying They Wouldn’t. Too Bad"

Great opportunity to join an amazing team!

Hi everyone, today we are announcing the start of the LibreFold project, a collaborative research workshop aiming to improve AlphaFold-like protein structure prediction methods. 1/8

Generative AI for protein design is now one click away. We've launched a new platform that puts powerful tools behind a simple web interface, making #proteindesign accessible to anyone. Try it here ➡️ huggingface.co/spaces/merle/P… Learn more on YouTube ➡️ youtu.be/bS71K2U0amA

The speed of progress is breathtaking. FrugalGPT by James Zou and colleagues contains lessons for scaling/improving any LLM.