Excited to give the keynote “Charting the Path from Modelling to Precision” at #CRINA2025 in Edmonton 🇨🇦!

Let’s talk about transparent, reproducible & frugal #AI for cancer research, the hallmarks of predictive oncology, and what’s next with agentic AI systems.

🔗 sites.google.com/ualberta.ca/...

Our goal is clear: build an AI-enabled institution where health care practitioners, researchers, administrators, and patients are supported by trustworthy AI ecosystems that truly improve outcomes.

Excited to work with exceptional colleagues across UHN to make this vision real!

I am honoured to step into the role of @uhn.ca
Executive AI Scientific Director and Co-Director of the
@uhnaihub.bsky.social

This is a major opportunity to accelerate how we bring responsible, transparent and effective AI into patient care, research, education and operations across UHN.

Excited to join the Challenge Me debate on Nov 6!

Motion: Method developers should be required to publish all limitations & failures.

I’ll argue against mandatory transparency, not because I oppose open science, but to stress-test the idea and push our thinking.

🗓️ Nov 6, 11am EDT
🔗 rb.gy/p71tnv

Thrilled to attend the Break Through Cancer Summit at Johns Hopkins as a SAB member.

The Data Science TeamLab is where AI and data science meet translational oncology, the sessions on (agentic) AI will chart new frontiers in how we cure cancer. 🚀

@break-cancer.bsky.social

Proud to share our latest work accelerating AI-driven precision oncology for rare T-cell leukemias & lymphomas.

We built TCL-38, the most comprehensive multi-omics + drug response atlas for TCL to predict selective drug sensitivities & combinations --> new therapeutic vulnerabilities & biomarkers.

🚨 Excited to chair the AI Applications in Human Health session at the MAQC 8th Annual Conference to make AI clinically relevant and reproducible.

Transparency, reproducibility & reusability aren’t optional—they’re the foundation for trustworthy AI in healthcare.

Join us: themaqc.org

#AI #MAQC25

Our new method GraphComm uses graph neural networks to decode cell–cell communication from single-cell RNA-seq @uhnresearch.ca‬
🔍 Predicts novel ligand–receptor interactions
🧠 Handles spatial data & drug perturbations
🌐 Open-source: github.com/bhklab/GraphComm

Full paper: doi.org/10.1038/s415...

How AI + open data can accelerate drug discovery?

New paper from @thesgc.bsky.social and led by UNC presents the first fully open AI framework for DNA-Encoded Library hit discovery, scanning billions of molecules and confirming novel binders.

👉 Data & models available on AIRCHECK aircheck.ai

Excited to join the Multi-AI for Omics Working Group of European Interdisciplinary Society for Artificial Intelligence in Cancer Research (ESAC) @esac-cancerai.bsky.social !
Today marks our inaugural meeting, bringing together global experts to shape the future of AI-driven cancer research.

Honoured to be featured in Open Access Government!
I shared how AI and biomarker discovery are transforming precision oncology, and why rethinking therapy response evaluation is essential to deliver truly personalized cancer care.
Read more 👉 edition.pagesuite-professional.co.uk/html5/reader...

Nice paper from Yu Li & Le Song on morphology prediction under perturbations.

Their method is called MorphDiff… which is super close to our MorphoDiff previously published at ICLR 2025.

Guess we’ll need a morphology diff checker now 😅

bsky.app/profile/bowa...
www.nature.com/articles/s41...

www.nature.com/articles/d41...
So let me get this straight
1. Academics spend decades building the PDB openly
2. A company trains AlphaFold on it, releases limited code under pressure
3. Academics work hard to create fully open versions
4. Companies then use the open models to build closed products

I will be presenting "Biomarkers of response to drugs and radiotherapy" in the AI Education track at #ESMO2023 in Madrid, Spain.

Here are the slides: bit.ly/3S9D6b1

They are open-source so feel free to re-use.

Any feedback is welcome!

That is to the EuroBioC2023 organizers, very nice in-person event at Ghent Universiteit. Here are my slides: Data and Packages For Meta-Analysis of Immunotherapy Clinical Trials in Cancer

bit.ly/47iIRsq

I will be giving a talk at the #EuroBioC2023 conference about our PredictioR Package For Meta-Analysis of Immunotherapy Clinical Trials in Cancer buff.ly/3MIdJsD

The package is based on our recent paper in Annals of Oncology:

doi.org/10.1016/j.an...