Leash Bio releases Minos, an agent sandbox built on DNA-encoded library data
Overview
Leash Bio releases Minos, an AI playground where agents reason about structure-activity relationships using lab-in-the-loop simulations derived from its DNA-encoded library data.
The company published a preprint, blog post and open dataset with the release, according to a Latent Space report.
Leash co-founder Alex Allman says DNA-encoded libraries generate billions of protein–molecule measurements for training AI drug-discovery agents. He says no team in Leash's Kaggle competition beat chance on an unseen chemical library, and that agents ran over 10,000 medicinal chemistry campaigns in a weekend in replayable experimental worlds.
Written by AI from the articles below · updated Oct 14, 5:07 PM ET
Check the sources:
Developments
2 developments
- Oct 14, 4:44 PM ET · 1 articleLeash Bio explains how DNA-encoded libraries could train AI drug-discovery agentsLatent.Space: Leash Bio explains how DNA-encoded libraries could train AI drug-discovery agents
- Oct 14, 4:35 PM ET · 1 articleLeash Bio releases Minos, an agent sandbox built on its DNA-encoded library dataLatent Space: Leash Bio releases Minos, an agent sandbox built on its DNA-encoded library data
Article timeline
The articles in this story. Times are ET.
Latent.Space@latentspacepodXLeash Bio explains how DNA-encoded libraries could train AI drug-discovery agentsAILeash Bio's Alex Allman explains how DNA-encoded libraries generate billions of protein–molecule measurements for AI training. He says no team in Leash's Kaggle competition beat chance on an unseen chemical library, and that agents ran over 10,000 medicinal chemistry campaigns in a weekend in replayable experimental worlds.

- Latent SpaceBlogLeash Bio releases Minos, an agent sandbox built on its DNA-encoded library data
AILeash Bio releases Minos, an AI playground where agents reason about structure-activity relationships using lab-in-the-loop simulations derived from its DNA-encoded library data. The company published a preprint, blog post, and open dataset alongside the release. The podcast episode with co-founder Ian Quigley covers Leash's DNA-encoded library work and its Kaggle competition results.
Heat trend
Not enough continuous observations to show a trend yet.