The goal of the 2024 machine learning challenge was to design algorithms that would enable a more accurate diagnosis of inflammatory bowel disease (IBD), affecting millions of people worldwide. The algorithms created for this three-part challenge offered a high-resolution view of IBD by integrating pathology images of inflamed gut tissue with spatial transcriptomics data. Genes were identified to serve as markers for potential cancerous regions in the gut, facilitating the early detection and treatment of colorectal cancer. The gene panels that were developed by top performers are currently being tested in patient samples through lab experiments at the Broad Institute!

Learn more about the Crunch 1 and Crunch 2 winners in our recap.

The challenge opened on October 28, 2024 with the following deadlines:

• Crunch 1: due February 7, 2025
• Crunch 2: due March 21, 2025
• Crunch 3: due April 18, 2025

Helpful links and information:

• PDF of the full specifications.
• Challenge page on Crunch. Find all competition rules, datasets, etc. here.
• Challenge overview video.
• Challenge introduction video.‍
• Lecture series for the biology, technology, and data background necessary to participate in the competition.

‍The challenge was offered as an IAP course in 2025 for MIT students:

• 6.S099 Machine Learning Challenge for Biomedical Discoveries
• Instructors: Caroline Uhler, Paul Blainey, Jonathan Weissman
• Schedule: Tuesday, Jan. 7 - Thursday, Jan. 30 (Tuesdays and Thursdays, 11:30 a.m. - 1 p.m., room 26-168 and on Zoom)