The Gevaert Lab has developed SEQUOIA, a transformer model that predicts cancer transcriptomic profiles from whole slide images, enabling cost-effective genetic insights from H&E images. It has been validated across multiple cancer types and shows potential in predicting cancer recurrence risk and spatial gene expression, supporting personalized cancer management. https://www.nature.com/articles/s41467-024-54182-5
The Zou Group is very excited to have a new paper on spatial single-cell aging clocks published in Nature, led by DBDS PhD student Eric Sun.
Read it here: https://www.nature.com/articles/s41586-024-08334-8
“Simple and effective embedding model for single-cell biology built from ChatGPT” from the Zou Group has been published in Nature Biomedical Engineering.
Read it here: https://www.nature.com/articles/s41551-024-01284-6
Doctors are using AI. Will it make healtchcare better or break it?
Millions are already being treated by doctors using artificial intelligence to take notes and draft emails to patients. But we still don’t know when AI tools are accurated or biased — or even if they save doctors time.
“I recognize the health-care system is broken. Access to care is a huge issue. Doctors make mistakes. I hope AI can solve that,” Daneshjou says.
Read the story here.