Why is progress against cancer so uneven? Because to make big gains we first had to understand the underlying biology💡 Identifying HER2 led to trastuzumab. Finding PD-1 led to pembrolizumab. And discovering a druggable pocket in a mutated form of KRAS, long considered “undruggable”, led to sotorasib. In each case it took decades from target discovery to drug on the market - most of that time went into understanding the disease. 🧠 At Owkin, we’re using AI to shorten that path, so the next breakthrough comes years sooner. Read more from the source - AACR Cancer Progress Report 2026: cancerprogressreport.aacr.or…

How do you find out which AI system is better at addressing your use-case? You use a benchmark. But using a benchmark isn’t always simple. A benchmark is a defined standard for comparing a model against other models, earlier versions, or human performance. For a task with a known answer, like classifying a scan, one score is often enough. For an AI scientist, which chooses its own analyses and interprets noisy results, benchmarking can require numerous metrics. For example, accuracy on a single run doesn’t necessarily tell you whether the system is reliable. So what you choose to measure matters as much as the score. Read our 5-minute guide to benchmarking in healthcare AI 👇 Benchmarking in Healthcare AI→ owkin.com/a-z-of-ai-for-heal… Read our blog on K Pro vs Claude → owkin.com/blog/how-does-k-pr…

In a recent interview on the @632nmPodcast, Nobel Prize-winning scientist Eric Betzig makes a point that applies well beyond microscopy: you can’t model what you haven’t observed. Proteins we thought formed stable complexes turned out to touch DNA for seconds. The same holds for AI in medicine. Models built on incomplete data give confident answers to the wrong questions. That’s why at §Owkin we start from real patients (real tissue, real outcomes, real validation) before we ask AI to discover anything. Worth the listen. youtube.com/watch?v=RUB37QhW…

That’s a wrap on BioTechX! In an ocean of blue & dark booths, suits and ties, we were hard to miss with our flashy white tee shirts and our lovely pink booth 🦩 Great conversations, inspiring keynotes (with maybe a few too many words per slide 😉), and lots of new connections. Thanks to everyone who stopped by, and huge kudos to our amazing crew: all smiles for hours, at the entrance by 8am handing out chocolates, and still there after closing pouring bubbles 🥂 Drug development doesn’t need more of the same. That’s why we’re building a state-of-the-art AI scientist, a true copilot for pharma and biotech teams, built on the world’s largest AI-ready real-world patient dataset, to accelerate drug development. Missed us at BioTechX? Don’t miss out twice. Book a demo now :) owkin.com/connect/book-a-dem…

Published in Genome Medicine: MOSAIC, our spatial multi-omics atlas of 2,700+ cancer samples. Plus MOSAIC Window, a first release of 60 patients across 5 tumour types with six matched data modalities, now on EGA. doi.org/10.1186/s13073-026-0…

When you train a model on two different data modalities, it can often perform better than when trained on just one. A review in Frontiers in Biomedical Technology this year, took stock of the field. The authors found that multimodal LLMs that combine imaging (MRI, CT, PET) with omics (genomics, transcriptomics, proteomics) allowed models to offer a holistic view of disease at structural and molecular levels. This helped support prognosis beyond what models based on one modality can offer. We’ve found the same. Our AI scientist - K Pro - has access to 6 modalities per patient through our MOSAIC dataset. This has allowed it to make predictions about cancer biology and suggest potential new therapeutic ADC molecules that wouldn’t be possible with single modality data. That’s the power of multimodal data. Read our 5-minute guide to multimodal data in healthcare AI → owkin.com/a-z-of-ai-for-heal… Read the review here → researchgate.net/publication…

Day 1 at BioTechX Europe 2026 is underway, and the Owkin team is on site in Basel! 🎉 Stop by Booth #178 to: 🔎 See how K Pro, our AI scientist for biology and pharma R&D, helps researchers tackle complex scientific questions and accelerate discovery 📊 Learn more about Owkin’s Patient Data Network, our partnership with leading healthcare organizations to expand access to high-quality, multimodal patient data for research. Early riser? We'll be outside the venue every morning of the conference handing out free pralines from Confiserie Bachmann, so stop by to say hello and start your day with something sweet 🍫 Visit us or book a meeting with our team via the link in the comments 👇 📍 Booth #178 | 📅 6-8 October 2026

How can 200 patients from Owkin’s datasets compare to 10,000 patients from others? The answer is that most large, real-world datasets weren’t built for biopharma R&D. They were built for care, billing or population research, and they possess stratospheric patient counts - but extremely low data density per patient. So they’re fragmented. Molecular tests exist when a clinician ordered them rather than on a protocol designed for data reuse, data modalities are limited to the standard of clinical care, and data analysis finds association rather than causality. Owkin’s data start’s with the patient’s treatment path. Baseline, adverse events, resistance, remission. Same block, same patient, same day. We work with our world-leading network of partner hospitals across the world to find multimodal data on the treatment path, find samples and cohorts that match the desired treatment path, run new assays on samples, and get it into research use at the speed of science. The reason you can do so much with 200 patients’ data from Owkin is because ours is collected specifically to answer the questions of biopharma R&D. Imagine having a comparator arm cohort in your chat window and your bioinformatics toolbox, across the entire asset lifecycle from target to corporate development. Find out more about how below 👇

Calling all hackers, builders and bio-experts 🧬💻 We’re partnering with @spc for a 48-hour Bio Data Hackathon on October 16-18 in San Francisco - and we’re providing access to a subset of our MOSAIC dataset, the world’s largest spatial multiomics dataset in oncology. Participants will have the opportunity to work with multimodal patient data from 60 patients across five cancer indications, spanning six different modalities: spatial transcriptomics, single-cell RNA-seq, bulk RNA-seq, whole exome sequencing, H&E pathology and clinical data. What new biological insights can you uncover when you bring these different layers of patient data together? From exploring biomarkers and patient subgroups to investigating treatment response and resistance mechanisms, we’re excited to see what participants build and discover. Whether you’re an ML or AI researcher, wet-lab biologist, bioinformatician, builder or founder, this is your chance to get hands-on with unique datasets and tools and build something new. Ready to hack biology? Applications close October 12. Apply now: southparkcommons.com/bio-dat…

🇪🇪 Today we announce a new collaboration with the country of Estonia, partnering on sovereign AI in biology. Working with Metrosert we'll evaluate how K Pro, Owkin’s AI scientist, can help approved researchers analyse complex Estonian health data inside secure national infrastructure, whilst always preserving data sovereignty. As Estonia’s Minister of Economic Affairs, @liinavahtras, puts it: “The next step is enabling those assets to create greater value for science and future health innovation.” That’s what this partnership is built to do. Read the full announcement: owkin.com/newsfeed/estonia-a…

What sets an AI Scientist apart from general-purpose AI? 🧪 Next week at @BioTechX_ Europe, our CTO Rodrigo Barnes will take the stage to explore how AI can support a new generation of pharma R&D. Drawing on examples from our current work, he’ll demonstrate how K Pro, Owkin's AI Scientist, helps pharma researchers move from scientific questions to testable hypotheses faster, and how these capabilities can scale to drive IP and asset development. 🎤 "Building an AI Scientist for a New Generation of Pharma R&D" 📅 7 October 2026 | 12:05 pm 📍 Theatre 9 If you’re attending, join us for the session to learn more. To see K Pro in action, stop by Booth #178 for a live demo or book a meeting with our team at BioTechX: owkin.com/our-events/biotech…

IMILIA is accepted at COMPAYL, MICCAI 2026. It predicts IBD inflammation from H&E slides and explains why: HistoPLUS quantifies the cells in the regions driving each prediction, and those counts alone are strong markers of inflammation. arxiv.org/abs/2512.13440

Still wondering which AI Scientist is right for you? Join the discussion this Thursday with Roy Granit (Compugen) and Jonas Beal (Owkin), as we explore the evolving AI Scientist landscape and what to consider when choosing one for your research. luma.com/khk5yc0e?utm_medium…

🎙️ Ulrich Keilholz, Director of Charité Comprehensive Cancer Center, joins VP of Product Jonas Béal on our podcast to discuss how we build expert biomedical skills w/leading academics and how Ulrich’s expertise in systems medicine can become an AI skill. youtu.be/PPfw78-SFvA

How does Roy Granit, Head of Computational Discovery at biotech Compugen, rate K Pro? Ahead of our webinar on 24th September, we asked him. Read the Q&A here → owkin.com/blog/why-compugen-… Sign up to the webinar → luma.com/khk5yc0e?utm_medium…

The AI scientist landscape is evolving rapidly, and choosing the right one for your research isn't always straightforward. Join us for an upcoming webinar with Roy Granit, Head of Computational Discovery at Compugen, and Jonas Béal, VP Product at Owkin. luma.com/khk5yc0e?utm_medium…