Christian Dallago

Christian Dallago

@machine.learning.bio

🏳️‍🌈 NVIDIA & Duke. Was Allianz, VantAI, TUM. BioCS+ML dude. Lab page: https://machine.learning.bio GScholar: https://scholar.google.com/citations?user=4q0fNGAAAAAJ

1.2K followers305 following61 posts

View full thread
Christian Dallago
@machine.learning.bio

A special shoutout to Riya Narain, who spent the summer interning on my team and did so much of the work behind this release. I’m proud of what she’s contributed.

View full thread
Christian Dallago
@machine.learning.bio

A huge team effort with Google DeepMind (Risha Patel), EMBL-EBI (Jennifer Fleming, Sameer Velankar) and SNU (Yewon Han, Martin Steinegger), University of Glasgow (Joe Grove), SIB (Philippe Le Mercier), CEPI (Newton Wahome) and Sungkyunkwan University (Milot Mirdita).

Christian Dallago
@machine.learning.bio

Building on the AlphaFold Database protein-complex release at GTC, we explored ~1.7M protein pairs across 2,800+ viral proteomes. That scale lets us compare how viral proteins assemble and identify conserved interfaces.

Christian Dallago reposted

EMBL-EBI
@ebi.embl.org

You asked, we listened. Millions of AI-predicted protein complex structures are now available in the #AlphaFold Database. This spans homodimers from 20 of the most studied species, including humans, as well as the World Health Organization’s priority pathogens list. www.ebi.ac.uk/about/news/t...

Christian Dallago
@machine.learning.bio

Five years ago, we released FLIP. The core question was: can ML models for protein fitness prediction generalize in the ways that actually matter for protein engineering, i.e. low data, extrapolation to more mutations, out-of-distribution sequences?

View full thread
Christian Dallago
@machine.learning.bio

Five years ago, we released FLIP. The core question was: can ML models for protein fitness prediction generalize in the ways that actually matter for protein engineering, i.e. low data, extrapolation to more mutations, out-of-distribution sequences?

View full thread
Christian Dallago
@machine.learning.bio

I love this gritty work here; there are no new architectures, no leaderboard-topping number to screenshot. However, it's how we (and hopefully the community) can measure whether the models we're all building and using are getting better where it counts — at the bench.

Christian Dallago
@machine.learning.bio

Five years ago, we released FLIP. The core question was: can ML models for protein fitness prediction generalize in the ways that actually matter for protein engineering, i.e. low data, extrapolation to more mutations, out-of-distribution sequences?

View full thread
Christian Dallago
@machine.learning.bio

The answer in 2026 is largely the same. Simple ridge regression on one-hot sequences, optionally supplemented with zero-shot pLM likelihoods, often matches or outperforms fine-tuned protein language models.

Christian Dallago
@machine.learning.bio

Five years ago, we released FLIP. The core question was: can ML models for protein fitness prediction generalize in the ways that actually matter for protein engineering, i.e. low data, extrapolation to more mutations, out-of-distribution sequences?

View full thread
Christian Dallago
@machine.learning.bio

We were interested in how things had changed 5 years after our first release. So, we built FLIP2 on select datasets from great labs across the world, many of which have gracefully agreed to make their data freely available.

Christian Dallago
@machine.learning.bio

FLIP2, adds seven new sequence-fitness landscapes - industrial enzymes, nucleases, rhodopsins, protein-protein interactions - and 16 splits that test the generalization axes protein engineers really hit: more mutations, new positions, higher fitness, different wild-types.

Christian Dallago
@machine.learning.bio

Five years ago, we released FLIP. The core question was: can ML models for protein fitness prediction generalize in the ways that actually matter for protein engineering, i.e. low data, extrapolation to more mutations, out-of-distribution sequences?

Christian Dallago
@machine.learning.bio

FLIP spawned fast development of several different benchmarking efforts across protein design, engineering, and variant effect assesment. The answer in 2021 was: sometimes, but simpler models hold up surprisingly well.

Christian Dallago reposted

Kevin K. Yang 楊凱筌
@kevinkaichuang.bsky.social

We made FLIP2, a protein fitness benchmark spanning seven new datasets, including enzymes, protein-protein interactions, and light-sensitive proteins, as well as splits that measure generalization relevant to real-world protein engineering campaigns.

Christian Dallago
@machine.learning.bio

Our latest protein family-based GenAI collection of tools and datasets, ProFam, is out now. Everything -- from data, training and inference code, to a 215M llama-based ProFam-1 are fully open sourced. 🧵

CATH-Gene3D@cathgene3d.bsky.social

Built by CATH, TÜM and NVIDIA, ProFam-1 is our new open-source protein family language model (pfLM) designed to generate functional protein variants and predict fitness using in-context example sequences.

View full thread
Christian Dallago
@machine.learning.bio

In essence, we probed the model's ability to ricapitulate family statistics, bootstrap protein structure prediction, and assess mutation effect, demonstrating excellent performance across all tasks, especially using test-time-scaling via prompt conditioning.

Christian Dallago
@machine.learning.bio

You can use the model right now to freely generate families for single sequence inputs (i.e., diversification conditioned by intrinsic representations of evolution), or to engineer proteins based on family promts (diversification by conditioning on particular evoluationary trajectories).

Christian Dallago
@machine.learning.bio

Our latest protein family-based GenAI collection of tools and datasets, ProFam, is out now. Everything -- from data, training and inference code, to a 215M llama-based ProFam-1 are fully open sourced. 🧵

CATH-Gene3D@cathgene3d.bsky.social

Built by CATH, TÜM and NVIDIA, ProFam-1 is our new open-source protein family language model (pfLM) designed to generate functional protein variants and predict fitness using in-context example sequences.

Christian Dallago
@machine.learning.bio

With ProFam-1, we scaled learning from single sequence to protein family definitions of different kinds, curating a large protein family corpus, ProtFam-atlas. I'm particularily stoked about the idea of inference-time-compute. This contribution laid out a very exciting path for future work.

Christian Dallago
@machine.learning.bio

Another exciting opportunity, this time as a colleague at Duke! Join as tenure track assistant prof. in Cell Bio & let’s work on closing the gap between in-silico and in-vivo: www.nature.com/naturecareer... Important: application closes Nov 1st!!!

nature.comTenure-Track Assistant Professor Position –AI/ML for Cell Biology - Durham, North Carolina (US) job with Duke University School of Medicine | 12844591Tenure-Track Assistant Professor Position –AI/ML for Cell Biology
Christian Dallago
@machine.learning.bio

Another opening: Senior Multiscale Biology Applied Research Scientist! nvidia.eightfold.ai/careers/job/... Are fascinated by fundamental data modalities across biology like RNA-seq, mass spec & want to build computational tools that harnessing data to build intelligence? Come: join the team!

nvidia.eightfold.aiSenior Applied Research Scientist, Multiscale Biology | NVIDIA CorporationApply your expertise in engineering biology through algorithms and tools for genes, tissues, organisms, and populations. Conduct collaborative applied research in multiscale biology using deep learnin...
Christian Dallago
@machine.learning.bio

Are you passionate about leading collaborative, fast moving, applied bioinformatics research projects that help the entire community move forward? Apply to work in my team at NVIDIA: nvidia.eightfold.ai/careers/job/...

nvidia.eightfold.aiSenior Applied Research Scientist, Bioinformatics | NVIDIA CorporationLead applied and collaborative research programs using bioinformatics, high performance computing, and deep learning for biological advancements. Develop and accelerate bioinformatics software and alg...

Christian Dallago reposted

Stephan Hacker
@stephanhacker2.bsky.social

Great talk by @machine.learning.bio at the 5th Virtual @chembiotalks.bsky.social. He talked about the use of #MachineLearning and #BigData to address biological questions. Cool insights into both predicting functions and designing proteins ieeexplore.ieee.org/document/947... arxiv.org/abs/2503.00710

ieeexplore.ieee.orgProtTrans: Toward Understanding the Language of Life Through Self-Supervised LearningComputational biology and bioinformatics provide vast data gold-mines from protein sequences, ideal for Language Models (LMs) taken from Natural Language Processing (NLP). These LMs reach for new pred...

Christian Dallago reposted

Nature Methods
@natmethods.nature.com

GPU-accelerated MMseqs2 offers tremendous speedup for homology retrieval, protein structure prediction with ColabFold, and protein structure search with Foldseek. @martinsteinegger.bsky.social @milot.bsky.social @machine.learning.bio www.nature.com/articles/s41...

nature.comGPU-accelerated homology search with MMseqs2 - Nature MethodsGraphics processing unit-accelerated MMseqs2 offers tremendous speedups for homology retrieval from metagenomic databases, query-centered multiple sequence alignment generation for structure predictio...

Christian Dallago reposted

Stephan Hacker
@stephanhacker2.bsky.social

Looking forward to hearing about the potential of machine learning for #Biology and #DrugDiscovery from an industry perspective. Register for the Virtual @chembiotalks.bsky.social to hear the perspective of Chris Dallago (@machine.learning.bio) from Nvidia. #ChemBio #Chemsky #ML #MachineLearning

Virtual ChemBioTalks@chembiotalks.bsky.social

The third talk of the 5th Virtual ChemBioTalks will be given by Chris Dallago (@machine.learning.bio). He will talk about the research at NVIDIA into “Industrial BioML: the coming of age of machine learning for biology”. Make sure to register for free: https://cvent.me/G1geWW

Christian Dallago reposted

César de la Fuente
@delafuentelab.bsky.social

(1/5) Venoms are a vast, largely untapped library of bioactive molecules—and our new paper in @natcomms.nature.com ‬ @natprot.nature.com reveals just how powerful they can be. 🐍⚡️

nature.comComputational exploration of global venoms for antimicrobial discovery with Venomics artificial intelligence - Nature CommunicationsResearchers used artificial intelligence to mine global venom proteomes and discovered novel peptides with antimicrobial activity. Several candidates showed efficacy against drug-resistant bacteria in...

Christian Dallago reposted

Stephan Hacker
@stephanhacker2.bsky.social

The final program is now online for the 5th Virtual @chembiotalks.bsky.social: web.cvent.com/event/60e9f3... Looking forward to talks by Sarah O'Connor, Chengqi Yi, @machine.learning.bio, @cathleenzeymer.bsky.social, @kellychibale.bsky.social, Jennifer Prescher and @craigmcrews.bsky.social.

Virtual ChemBioTalks@chembiotalks.bsky.social

The registration link and program for the 5th Virtual ChemBioTalks are finally available! If you are interested in the field of Chemical Biology, make sure to register for this free, virtual event on September 30th, 2025! cvent.me/G1geWW (1/4) #Chemsky #ChemBio