How to learn gene regulation patterns from multimodal single-cell data? To answer this question, we at OpenProblems organised the Kaggle competition on modality prediction. It ran back in 2022 but still remains the world's largest competition in the single-cell field. 1/17
26.02.2026 20:34 — 👍 5 🔁 2 💬 1 📌 0
At the 132nd Internat. Titisee Conference on Biology 2.0: The AI Revolution in Biology & Medicine
From sequence→function models 🧬
to protein & generative structure models 🧪
to AI of cell states & perturbations 🧫
Great science, great friends, beautiful lake. Thanks @BIFonds!
Join us at the #CellPressSymposia: Decoding cellular complexity: Single cell biology in the era of AI, Dec 2–4, 2026, Munich, Germany, in partnership with @helmholtzmunich.bsky.social.
#CPSsinglecellbiology2026 @ethz.ch.bsky.social
http://dlvr.it/TR8Rw9
Join us at Cell Symposia: Single-Cell Biology in the Era of AI (Dec 2–4, 2026, Munich) 🇩🇪
From multi-omics to AI-driven models of cellular dynamics — an exciting lineup across experimental & computational biology.
Hope to see you there!
🔗 www.cell-symposia.com/single-cell-...
Looking forward to tonight’s Synapse Munich fireside chat 🔥
Great to connect with the Munich AI x biotech community. Thanks to Nucleate and Amino Collective for the invitation! See you there 🚀
luma.com/ehl0ap32
We discuss anchoring strategies, generative transport, cross-modality attention, and integrating priors (GRNs, ontologies, pathways, spatial context).
Compositionality may also help bridge scales → toward “virtual cell” systems 🤖
Thanks to Mojtaba Bahrami for leading this 🙌
Core idea: build multimodal single-cell FMs modularly, combine expert models (RNA, ATAC, proteomics, imaging, text) via attention interfaces.
Why? Paired multimodal data are rare.
Unimodal data are abundant.
Modalities are often complementary, not redundant 🧬
🚀 From modality-specific to compositional foundation models for cell biology
Happy to share our new Perspective in Cell Systems.
We describe an emerging shift toward compositional multimodal architectures 🧩
authors.elsevier.com/c/1mdzg8YyDf...
Munich takes the top two spots in the FT ranking of Europe’s startup hubs — UnternehmerTUM #1, Start2 Group #2.
Strong science, engineering depth, & real bridges between research + entrepreneurship. Huge opportunity ahead in part. for AI + bio/med. 🚀
www.ft.com/content/ab05...
Honored to receive today the official membership document of the @Leopoldina – Germany’s National Academy of Sciences – in Halle.
“Nunquam otiosus” – never idle. A reminder that science is a continuous responsibility.
And proud to be elected together with my AI & ELLIS colleague Sepp Hochreiter.
Looking forward to “Europe’s Fight for BioTech.”
AI is no longer a tool around biotech - it’s becoming the core engine. Pharma is turning into techbio: data-driven, model-based, AI-native.
Biotech = software-defined biology.
Can Europe move fast enough?
www.unternehmertum.de/en/landingpa...
Where tech meets politics.
At the #MSC Startup Hub, innovation & geopolitics collide in real time. Our @Helmholtz_de President Martin Keller is chairing session with former MI6, FBI/Interpol & Ukrainian security leader. Clearly security, crypto & AI are strategic capabilities.
🚀 New preprint: population-aware single-cell dynamics
Collab w/ Bertie Göttgens’ lab, led by Weizhong.
pseudodynamics+ goes beyond pseudotime to infer birth/death, differentiation & interpolate time in scalable fashion - directly from snapshot scRNA-seq.
👉 www.biorxiv.org/content/10.6...
CellRank protocol now in @NatureProtocols! 🎉
@PhilippWeiler's detailed guide shows how to:
- Integrate velocity, pseudotime & time points
- Infer cell fate probabilities
- Analyze lineage priming at scale
Great complement to our CellRank 2 paper!
📖 www.nature.com/articles/s41...
Did you catch our recent Protocol describing the inference of cell fate from single-cell RNA sequencing data using CellRank?
🧪 @natprot.nature.com @fabiantheis.bsky.social
A recent @derspiegel article on “one of Germany’s most renowned AI researchers” got me thinking. The person was new to me, but I think highlights a somewhat recurring issue: media visibility is not the same as scientific contribution.
www.spiegel.de/wissenschaft...
🚀 Interested in a postdoc in computational biology & AI for neuroscience?
The NOMIS Synergy Fellowship offers a great opportunity to develop such a position in my lab within the interdisciplinary Synergy excellence cluster.
🔗 www.synergy-munich.de/news-events/...
Congratulations to Matthias Tschöp on his inauguration as President of @LMU_Muenchen! A fitting moment to celebrate 200 years of LMU and to reflect on the responsibility and impact of universities & science in society.
03.02.2026 17:26 — 👍 0 🔁 0 💬 0 📌 0
🚀 Excited to share new paper in Nature Genetics: Human Retina Cell Atlas (HRCA) with ~4M cells, >130 retinal cell types, deep scRNA/ATAC integration. Enables cross-species comparisons & GWAS cell-type enrichment. Huge thanks to Ignacio Ibarra & Chen lab.
www.nature.com/articles/s41...
Exciting Health AI session at #DLD2026 in Munich - with discussions about leveraging AI to integrate scales and design next experiments and treatments! A la DLD motto - it’s gonna get wild - follow at /www.dld-conference.com
16.01.2026 09:20 — 👍 3 🔁 0 💬 0 📌 0
A wonderful evening with my wife Michaela at the New Year’s Reception of the Bavarian Government ✨🥂
11.01.2026 16:33 — 👍 7 🔁 0 💬 0 📌 0
🔬 pertpy: a unified, scalable framework for single-cell perturbation analysis, now out in Nature Methods
Designed for modern perturbation data - CRISPR, drug screens, patient treatments - scaling to millions of cells and 1000s of conditions.
👉 www.nature.com/articles/s41...
🔬 pertpy: a unified, scalable framework for single-cell perturbation analysis, now out in Nature Methods
Designed for modern perturbation data - CRISPR, drug screens, patient treatments - scaling to millions of cells and 1000s of conditions.
👉 www.nature.com/articles/s41...
Thanks to team for wonderful lab xmas party yesterday! 🎄🎅🏼 Besides Feuerzangenbowle (=mulled wine^2) we had this funny guess-the-paper from Santa's crazy review. Not sure you can guess the paper 📝, but I loved ~summary "Benchmarking the benchmark of benchmarking benchmarks."😅
19.12.2025 19:28 — 👍 6 🔁 0 💬 0 📌 0
🎄 We’re hiring in Health AI at @HelmholtzMunich!
🔹 PIs in Health AI/AI Foundations
🔹 Health AI Fellows
🔹 Clinical AI Consultants
Apply👇
PIs: jobs.helmholtz-muenchen.de/jobposting/8...
Fellows: jobs.helmholtz-muenchen.de/jobposting/7...
Consult: jobs.helmholtz-muenchen.de/jobposting/c...
🚀 New preprint!
The Human Cytokine Dictionary - 9.7M cells, 12 donors, 90 cytokines - our largest human single-cell perturbation atlas yet. Users can map cytokine activity in their own data. Huge thanks to Lukas, Sören, Larsen, Parse Bio & Seelig lab!
🔗 www.biorxiv.org/content/10.6...
Nature Methods names AI agents a “Method to Watch”!
We've started building active-learning loops with perturbation models, where AI proposes experiments & data closes loop. Supplying such models to agents could transform exp design into theory search.
www.nature.com/articles/s41...
Excited to share new preprint led by Sergio Salas on “unassigned RNAs” in imaging-based spatial transcriptomics. Across technologies, 30-40% of transcripts fall outside segmentation, & many show structured, biologically meaningful extrasomatic patterns.
📄 www.biorxiv.org/content/10.6...
🚀 Excited to share our preprint scConcept: a contrastive pretraining framework for technology-agnostic single-cell representations.
We move beyond reconstruction to learn rich, cell-level embeddings for diverse omics.
📘 www.biorxiv.org/content/10.1...
Breaking down walls in #PrecisionHealth at #FallingWalls 2025! Great exchange with Frank Laukien, Caroline Gubser-Keller, Eicke Latz & Steve Quake on how AI & molecular analytics reshape medicine.
Our goal: virtual twin cells, tissues & patients.
#AIinMedicine #HelmholtzMunich
