Spatial transcriptomics holds enormous potential in oncology research by mapping gene expression within spatial tissue context. But generating this data is expensive, slow, and not always scalable.
To tackle this challenge, the Global AI Hackathon 2025 by Elucidata invited data scientists and ML practitioners worldwide to build AI models that could predict tumor gene expression directly from histology (H&E) slides.
We'll be marking the finish line of the Global AI Hackathon with a Symposium on Spatial Transcriptomics.
.webp)
Lorem ipsum dolor sit amet consectetur. Dictumst faucibus nibh imperdiet phasellus vitae ut sit. Ut eros amet massa tellus orci. Vestibulum ac arcu est nulla non eget nulla. Eget pulvinar eu ac mi cursus elementum neque. Massa nisl fringilla platea diam faucibus nullam. In lacus mauris nec ultrices. Ut accumsan leo adipiscing montes proin.
In this keynote session, Dr Ma'ayan delves into the transformative impact of generative AI and LLMs in the analysis of gene sets. He discusses how these innovations are enhancing gene annotation prediction, ultimately leading to the discovery of novel therapeutic targets across a broad spectrum of biomedical research applications.
Lorem ipsum dolor sit amet consectetur. Dictumst faucibus nibh imperdiet phasellus vitae ut sit. Ut eros amet massa tellus orci. Vestibulum ac arcu est nulla non eget nulla. Eget pulvinar eu ac mi cursus elementum neque. Massa nisl fringilla platea diam faucibus nullam. In lacus mauris nec ultrices. Ut accumsan leo adipiscing montes proin.
In this keynote session, Dr Ma'ayan delves into the transformative impact of generative AI and LLMs in the analysis of gene sets. He discusses how these innovations are enhancing gene annotation prediction, ultimately leading to the discovery of novel therapeutic targets across a broad spectrum of biomedical research applications.
Top 3 teams present their novel ML pipeline to predict the spatially resolved cell-type composition of a tissue sample using only its H&E slide image
%201.png)
Techniques that actually moved the needle: pre-processing, feature engineering, loss functions, and more.
.svg)
.svg)
.svg)
%201%20(1).svg)
%20(1).svg)
%201%20(1).svg)
.svg)
%201.svg){kind=icon}
.webp)
