CM4AI  ¡Ñ  AnVIL
DREAM CHALLENGE ¡P 2026¡V2027
CELL MAP PREDICTION DREAM CHALLENGE

Map the cell.
Advance discovery.

CM4AI¡VAnVIL Cell Map Prediction
DREAM Challenge

AnVIL and CM4AI invite individual researchers and teams to help advance in silico microscopy by developing AI models that predict where proteins localize within human cells.

The challenge

A human cell contains thousands of different proteins, but current imaging methods can visualize only a fraction of them at once. The Cell Map Prediction DREAM Challenge invites individuals and teams to help bridge that gap.

Using reference microscopy images and target protein amino acid sequences, participants will build models that predict protein localization and generate corresponding immunofluorescence microscopy images. A starter kit of training data will be available through NHGRI AnVIL, the cloud platform supporting reproducible model development and submissions.
ANVIL AT THE HEART OF THE CHALLENGE
Data, cloud computing, and reproducible research

The NHGRI AnVIL platform will provide access to starter training data and support reproducible workflows for models evaluated on new datasets. Explore the challenge and learn how AnVIL helps make large-scale, collaborative science possible.

Explore the challenge on AnVIL ¡÷
CHALLENGE TIMELINE
OPENS
October 19, 2026
CLOSES
February 1, 2027

$120,000 in awards & credits

Top-performing submissions will be eligible for cash prizes and Google Cloud computing credits, along with co-authorship opportunities in the overall challenge paper.

PLACEMENT
CASH
CLOUD CREDITS
1st place
$25,000
$40,000
2nd place
$15,000
$20,000
3rd place
$10,000
$10,000

Join as an individual or a team

Whether you are participating on your own or collaborating with a team, pre-register to receive challenge updates and explore the competition details.

Read the AnVIL event announcement  ¡P  View the challenge flyer

Featuring the NHGRI AnVIL cloud platform, with Cell Maps for AI (CM4AI), Bridge2AI, and the DREAM Challenge community.

Learn more at anvilproject.org