Julien Vibert (李满山)
@julienmanshan.bsky.social
MD, PhD | Medical Oncologist & Postdoctoral Researcher | AI & Computational Oncology | Sarcoma & Rare Cancers | Gustave Roussy & Heidelberg | Kung Fu practitioner | Philosophy lover |
Reposted by Julien Vibert (李满山)
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.
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.
November 11, 2025 at 11:14 PM
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.
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.
Reposted by Julien Vibert (李满山)
Rare cancers make up 22% of adult cases but cause 30% of cancer deaths.
At EORTC, we fight for patients with rare & ultra-rare cancers, creating trials that set new standards of care.
Help us keep hope alive.
🔗 Donate: donate.eortcresearchfund.org/eoy/
#NoPatientLeftBehind
At EORTC, we fight for patients with rare & ultra-rare cancers, creating trials that set new standards of care.
Help us keep hope alive.
🔗 Donate: donate.eortcresearchfund.org/eoy/
#NoPatientLeftBehind
November 18, 2025 at 10:53 AM
Rare cancers make up 22% of adult cases but cause 30% of cancer deaths.
At EORTC, we fight for patients with rare & ultra-rare cancers, creating trials that set new standards of care.
Help us keep hope alive.
🔗 Donate: donate.eortcresearchfund.org/eoy/
#NoPatientLeftBehind
At EORTC, we fight for patients with rare & ultra-rare cancers, creating trials that set new standards of care.
Help us keep hope alive.
🔗 Donate: donate.eortcresearchfund.org/eoy/
#NoPatientLeftBehind
Reposted by Julien Vibert (李满山)
📢 @fionakolbinger.bsky.social and @jnkt.bsky.social characterize the discordance between metrics used to evaluate AI tools and their clinical impact, proposing a framework to reduce this disconnect. www.nature.com/articles/s43... @tudresden.bsky.social @ekfzdigital.bsky.social
🔓 rdcu.be/eQmQp
🔓 rdcu.be/eQmQp
Adaptive validation strategies for real-world clinical artificial intelligence - Nature Computational Science
Technical metrics used to evaluate medical artificial intelligence tools often fail to predict their clinical impact. We characterize this discordance and propose a framework of study designs to guide...
www.nature.com
November 17, 2025 at 3:23 PM
📢 @fionakolbinger.bsky.social and @jnkt.bsky.social characterize the discordance between metrics used to evaluate AI tools and their clinical impact, proposing a framework to reduce this disconnect. www.nature.com/articles/s43... @tudresden.bsky.social @ekfzdigital.bsky.social
🔓 rdcu.be/eQmQp
🔓 rdcu.be/eQmQp
Reposted by Julien Vibert (李满山)
Register for the next EuSARC conference in Oslo, Norway, in June 2026
oslocancercluster.no/events/8th-e...
oslocancercluster.no/events/8th-e...
November 15, 2025 at 8:26 PM
Register for the next EuSARC conference in Oslo, Norway, in June 2026
oslocancercluster.no/events/8th-e...
oslocancercluster.no/events/8th-e...
Reposted by Julien Vibert (李满山)
Brian lived a life driven by compassion, service, & deep impact on the global #sarcomacommunity. We will honor him by staying steadfast in our commitment to research, collaboration, and supporting every patient and family affected by this disease.
November 15, 2025 at 8:17 PM
Brian lived a life driven by compassion, service, & deep impact on the global #sarcomacommunity. We will honor him by staying steadfast in our commitment to research, collaboration, and supporting every patient and family affected by this disease.
Reposted by Julien Vibert (李满山)
🚀 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...
We move beyond reconstruction to learn rich, cell-level embeddings for diverse omics.
📘 www.biorxiv.org/content/10.1...
scConcept: Contrastive pretraining for technology-agnostic single-cell representations beyond reconstruction
Recent large-scale single-cell foundation models have shown promise for exploring cellular states, yet they often underperform compared to simpler, domain-specific methods, raising concerns about…
www.biorxiv.org
November 13, 2025 at 5:17 PM
🚀 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...
We move beyond reconstruction to learn rich, cell-level embeddings for diverse omics.
📘 www.biorxiv.org/content/10.1...
Reposted by Julien Vibert (李满山)
Reposted by Julien Vibert (李满山)
Multimodal learning enables chat-based exploration of single-cell data - @bocklab.bsky.social @cemm.oeaw.ac.at @meduniwien.ac.at go.nature.com/3WPGJnW
Multimodal learning enables chat-based exploration of single-cell data - Nature Biotechnology
CellWhisperer uses multimodal learning of transcriptomes and text to answer questions about single-cell RNA-sequencing data.
go.nature.com
November 11, 2025 at 4:58 PM
Multimodal learning enables chat-based exploration of single-cell data - @bocklab.bsky.social @cemm.oeaw.ac.at @meduniwien.ac.at go.nature.com/3WPGJnW
Reposted by Julien Vibert (李满山)
November 11, 2025 at 5:49 PM
Reposted by Julien Vibert (李满山)
#PEAKtrial in #GIST positive as it could be.. great news for parients!! Congrats to @cogent
mPFS of 16.5 months is very impressive. Topline suggests new standard of care!!
@saxotomy.bsky.social @herbloongmd.bsky.social and @ all other folks out there
mPFS of 16.5 months is very impressive. Topline suggests new standard of care!!
@saxotomy.bsky.social @herbloongmd.bsky.social and @ all other folks out there
November 10, 2025 at 2:19 PM
#PEAKtrial in #GIST positive as it could be.. great news for parients!! Congrats to @cogent
mPFS of 16.5 months is very impressive. Topline suggests new standard of care!!
@saxotomy.bsky.social @herbloongmd.bsky.social and @ all other folks out there
mPFS of 16.5 months is very impressive. Topline suggests new standard of care!!
@saxotomy.bsky.social @herbloongmd.bsky.social and @ all other folks out there
Reposted by Julien Vibert (李满山)
I'm impressed by how far this model class has taken us—MOFA-FLEX may be the final iteration—a framework geared more towards generating models than being a standalone model itself. User-friendly, probabilistic programming rocks! We look forward to your feedback and experiences trying it out!
Are you using any of our factor models, such as MOFA? 🛵
You might’ve found it challenging to tailor them to your specific use cases - not anymore!
Introducing MOFA-FLEX: a flexible, modular factor analysis framework designed for customizable modeling across diverse multi-omics data scenarios. 1/n
You might’ve found it challenging to tailor them to your specific use cases - not anymore!
Introducing MOFA-FLEX: a flexible, modular factor analysis framework designed for customizable modeling across diverse multi-omics data scenarios. 1/n
November 8, 2025 at 9:06 PM
I'm impressed by how far this model class has taken us—MOFA-FLEX may be the final iteration—a framework geared more towards generating models than being a standalone model itself. User-friendly, probabilistic programming rocks! We look forward to your feedback and experiences trying it out!
Reposted by Julien Vibert (李满山)
🎗️ Attending #CTOS2025? Don’t miss the Foundational Science of Sarcomas Symposium!
This year’s highlights include sessions on
🔹 Cancer Metabolism & Drug Development
🔹 AI Applications in Sarcoma Diagnostics & Care
🔹 Publishing Your Research — In Conversation with Editors of Leading Oncology Journals
This year’s highlights include sessions on
🔹 Cancer Metabolism & Drug Development
🔹 AI Applications in Sarcoma Diagnostics & Care
🔹 Publishing Your Research — In Conversation with Editors of Leading Oncology Journals
November 7, 2025 at 3:20 PM
🎗️ Attending #CTOS2025? Don’t miss the Foundational Science of Sarcomas Symposium!
This year’s highlights include sessions on
🔹 Cancer Metabolism & Drug Development
🔹 AI Applications in Sarcoma Diagnostics & Care
🔹 Publishing Your Research — In Conversation with Editors of Leading Oncology Journals
This year’s highlights include sessions on
🔹 Cancer Metabolism & Drug Development
🔹 AI Applications in Sarcoma Diagnostics & Care
🔹 Publishing Your Research — In Conversation with Editors of Leading Oncology Journals
Reposted by Julien Vibert (李满山)
🚀 Prophet v3 is out!
AI that predicts how cells respond to genetic & chemical perturbations across readouts.
Now scales to 1.9 M molecules + in vitro validation of melanoma-specific hits 🧬💊
📄 biorxiv.org/content/10.1...
#AI #Biology #DeepLearning
AI that predicts how cells respond to genetic & chemical perturbations across readouts.
Now scales to 1.9 M molecules + in vitro validation of melanoma-specific hits 🧬💊
📄 biorxiv.org/content/10.1...
#AI #Biology #DeepLearning
November 4, 2025 at 5:40 PM
🚀 Prophet v3 is out!
AI that predicts how cells respond to genetic & chemical perturbations across readouts.
Now scales to 1.9 M molecules + in vitro validation of melanoma-specific hits 🧬💊
📄 biorxiv.org/content/10.1...
#AI #Biology #DeepLearning
AI that predicts how cells respond to genetic & chemical perturbations across readouts.
Now scales to 1.9 M molecules + in vitro validation of melanoma-specific hits 🧬💊
📄 biorxiv.org/content/10.1...
#AI #Biology #DeepLearning
Reposted by Julien Vibert (李满山)
Large language models (LLMs) may not reliably acknowledge a user’s incorrect beliefs, according to a paper in Nature Machine Intelligence. The findings highlight the need for careful use of LLM outputs in high-stakes decisions. go.nature.com/48VRpIQ 🧪
November 3, 2025 at 8:14 PM
Large language models (LLMs) may not reliably acknowledge a user’s incorrect beliefs, according to a paper in Nature Machine Intelligence. The findings highlight the need for careful use of LLM outputs in high-stakes decisions. go.nature.com/48VRpIQ 🧪
Reposted by Julien Vibert (李满山)
Join the advanced international course on
"Innovative approaches for pediatric cancer immunotherapy"
at the @institutcurie.bsky.social
January 28th - 30th , 2026
training.institut-curie.org/courses/inno...
"Innovative approaches for pediatric cancer immunotherapy"
at the @institutcurie.bsky.social
January 28th - 30th , 2026
training.institut-curie.org/courses/inno...
Innovative approaches for pediatric cancer immunotherapy | Institut Curie Advanced Training
This course aims to synthesize the state-of-the-art and the therapeutic perspectives related to harnessing children's immune system to target pediatric tumors. Encompassing topics from biological char...
training.institut-curie.org
November 2, 2025 at 7:10 AM
Join the advanced international course on
"Innovative approaches for pediatric cancer immunotherapy"
at the @institutcurie.bsky.social
January 28th - 30th , 2026
training.institut-curie.org/courses/inno...
"Innovative approaches for pediatric cancer immunotherapy"
at the @institutcurie.bsky.social
January 28th - 30th , 2026
training.institut-curie.org/courses/inno...
Reposted by Julien Vibert (李满山)
🧬 Excited to share Nicheformer out now in Nature Methods!
A transformer foundation model linking single-cell & spatial omics, learning spatial context from gene expression to map tissue organization.
Led by Ale Tejada & Anna Schaar 👏
👉 www.nature.com/articles/s41...
A transformer foundation model linking single-cell & spatial omics, learning spatial context from gene expression to map tissue organization.
Led by Ale Tejada & Anna Schaar 👏
👉 www.nature.com/articles/s41...
October 30, 2025 at 7:19 PM
🧬 Excited to share Nicheformer out now in Nature Methods!
A transformer foundation model linking single-cell & spatial omics, learning spatial context from gene expression to map tissue organization.
Led by Ale Tejada & Anna Schaar 👏
👉 www.nature.com/articles/s41...
A transformer foundation model linking single-cell & spatial omics, learning spatial context from gene expression to map tissue organization.
Led by Ale Tejada & Anna Schaar 👏
👉 www.nature.com/articles/s41...
Reposted by Julien Vibert (李满山)
Chromatin organization is measured in single cells using droplet microfluidics go.nature.com/3NCqwO1
rdcu.be/eNah5
rdcu.be/eNah5
Droplet Hi-C enables scalable, single-cell profiling of chromatin architecture in heterogeneous tissues - Nature Biotechnology
Chromatin organization is measured in single cells using droplet microfluidics.
go.nature.com
October 30, 2025 at 1:36 AM
Chromatin organization is measured in single cells using droplet microfluidics go.nature.com/3NCqwO1
rdcu.be/eNah5
rdcu.be/eNah5
Reposted by Julien Vibert (李满山)
Reposted by Julien Vibert (李满山)
excited to host this great lineup of speakers on the future of #pediatriccancer #immunotherapy @institutcurie.bsky.social @gustaveroussy.fr @ap-hp.bsky.social #ParisKidsCancer
October 27, 2025 at 2:58 PM
excited to host this great lineup of speakers on the future of #pediatriccancer #immunotherapy @institutcurie.bsky.social @gustaveroussy.fr @ap-hp.bsky.social #ParisKidsCancer
Reposted by Julien Vibert (李满山)
🚀 Our new Science paper is out (w/ B DeMeo, D Burkhardt, A Shalek, M Cortes): www.science.org/doi/10.1126/...
We show that active learning + transcriptomic perturbations can guide which exps to run next, boosting phenotypic hit rates >13x. AI not just predicting bio, but designing it. 🔁
We show that active learning + transcriptomic perturbations can guide which exps to run next, boosting phenotypic hit rates >13x. AI not just predicting bio, but designing it. 🔁
October 25, 2025 at 1:24 PM
🚀 Our new Science paper is out (w/ B DeMeo, D Burkhardt, A Shalek, M Cortes): www.science.org/doi/10.1126/...
We show that active learning + transcriptomic perturbations can guide which exps to run next, boosting phenotypic hit rates >13x. AI not just predicting bio, but designing it. 🔁
We show that active learning + transcriptomic perturbations can guide which exps to run next, boosting phenotypic hit rates >13x. AI not just predicting bio, but designing it. 🔁
This week I had the pleasure of co-chairing the Proffered Papers session on AI and Digital Oncology at ESMO 2025 in Berlin, together with Prof. Jakob Kather @jnkt.bsky.social, a leading figure in AI for oncology and my future postdoctoral mentor.
October 25, 2025 at 2:31 PM
This week I had the pleasure of co-chairing the Proffered Papers session on AI and Digital Oncology at ESMO 2025 in Berlin, together with Prof. Jakob Kather @jnkt.bsky.social, a leading figure in AI for oncology and my future postdoctoral mentor.
Reposted by Julien Vibert (李满山)
A new study introduces a publicly available script concordance test benchmark that reveals important limitations in the clinical reasoning capabilities of leading LLMs, particularly in processing new information under uncertainty. nejm.ai/3KlD5yX
@liamgmccoy.bsky.social #AI #MedSky #MLSky
@liamgmccoy.bsky.social #AI #MedSky #MLSky
October 24, 2025 at 1:15 PM
A new study introduces a publicly available script concordance test benchmark that reveals important limitations in the clinical reasoning capabilities of leading LLMs, particularly in processing new information under uncertainty. nejm.ai/3KlD5yX
@liamgmccoy.bsky.social #AI #MedSky #MLSky
@liamgmccoy.bsky.social #AI #MedSky #MLSky
Reposted by Julien Vibert (李满山)
Active learning with DrugReflector beats SotA in phenotypic hit-rate for virtual screening. Includes a sc perturbation dataset with 10 lines and 104 compounds. Out in @science.org now!
Grateful to Cellarity and @fabiantheis.bsky.social for the opportunity to contribute to this outstanding project!
Grateful to Cellarity and @fabiantheis.bsky.social for the opportunity to contribute to this outstanding project!
Active learning framework leveraging transcriptomics identifies modulators of disease phenotypes
Phenotypic drug screening remains constrained by the vastness of chemical space and technical challenges scaling experimental workflows. To overcome these barriers, computational methods have been dev...
www.science.org
October 23, 2025 at 7:41 PM
Active learning with DrugReflector beats SotA in phenotypic hit-rate for virtual screening. Includes a sc perturbation dataset with 10 lines and 104 compounds. Out in @science.org now!
Grateful to Cellarity and @fabiantheis.bsky.social for the opportunity to contribute to this outstanding project!
Grateful to Cellarity and @fabiantheis.bsky.social for the opportunity to contribute to this outstanding project!
Reposted by Julien Vibert (李满山)
scooby achieves DNA sequence-based single-cell level modeling of RNA-seq coverage and ATAC-seq insertion profiles by adapting a deep learning model that predicts bulk RNA-seq coverage. @johahi.bsky.social @lauradmartens.bsky.social @gagneurlab.bsky.social @tum.de
www.nature.com/articles/s41...
www.nature.com/articles/s41...
scooby: modeling multimodal genomic profiles from DNA sequence at single-cell resolution - Nature Methods
scooby achieves DNA sequence-based single-cell level modeling of RNA-sequencing coverage and ATAC-sequencing insertion profiles by adapting a deep learning model that predicts bulk RNA-sequencing cove...
www.nature.com
October 22, 2025 at 4:48 PM
scooby achieves DNA sequence-based single-cell level modeling of RNA-seq coverage and ATAC-seq insertion profiles by adapting a deep learning model that predicts bulk RNA-seq coverage. @johahi.bsky.social @lauradmartens.bsky.social @gagneurlab.bsky.social @tum.de
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Reposted by Julien Vibert (李满山)
The ESMO Guidance on the use of #LargeLanguageModels in #ClinicalPractice (#ELCAP) is out: the first structured set of recommendations to bring #AI language models into #oncology safely and effectively.
#ESMOAI25 #ESMO25 #LLMs #HealthAI
🔗 buff.ly/pg1gkiU
#ESMOAI25 #ESMO25 #LLMs #HealthAI
🔗 buff.ly/pg1gkiU
October 19, 2025 at 7:45 AM
The ESMO Guidance on the use of #LargeLanguageModels in #ClinicalPractice (#ELCAP) is out: the first structured set of recommendations to bring #AI language models into #oncology safely and effectively.
#ESMOAI25 #ESMO25 #LLMs #HealthAI
🔗 buff.ly/pg1gkiU
#ESMOAI25 #ESMO25 #LLMs #HealthAI
🔗 buff.ly/pg1gkiU