We’re running it back.

After our first agentic bio hackathon, we’re hosting a second - this time with real wet-lab validation, in collaboration with Adaptyv.

Teams will design protein binders for TREM2, a microglial receptor implicated in Alzheimer’s and related neurodegenerative diseases. Top computational designs will be experimentally tested and shared publicly.

Who this is for

Bench scientists who want to push on real targets with state-of-the-art tools
Computational biologists and protein designers
Engineers curious about biology
Anyone interested in agentic systems applied to real science

Expect live leaderboards, actually good coffee, food, and generous compute support from Rowan, OpenRouter, and Modal.

Note: this is in-person and will be in San Francisco - exact location will be updated soon.

What you’ll do

You’ll work in teams to:

Use bioArena to orchestrate agents and scientific tools
Design binders targeting TREM2
Submit your top candidates (max. 10) for scoring and ranking

In parallel, autonomous agents will be solving the same problem - a friendly competition to see how human-in-the-loop designs stack up against bots when the biology actually matters.

How scoring will work

Teams submit up to 10 binders, which are then ranked using ipSAE. A total of 100 binders will be sent to Adaptyv for wet-lab validation, selected from top-ranked designs, judge picks, and autonomous agent submissions.

The top three teams will be selected as winners based on experimental binding performance measured in Adaptyv’s binding assays, including binding classification and affinity metrics (e.g., KD).

May the best binders win.

bioArena Hackathon In Silico to In Vitro