Leash Bio Unveils Minos Pretrained Model and 6.5M-Molecule DEL Dataset
Current overview (not historical)
Leash Bio’s Ian Quigley and team have released the Minos pretrained model alongside a new DEL (DNA‑Encoded Library) dataset containing 6.5 million molecules. The dataset is built from high‑quality protein‑small‑molecule interaction data collected using DNA‑encoding library technology, and the authors claim it can supply tens of thousands of “lab‑level” simulations to address the industry’s data‑quality challenges. The release is positioned as a tool to train reinforcement‑learning models on DEL data, potentially accelerating scientific discovery. While the dataset’s scale and claimed simulation fidelity are highlighted, independent verification of the quality and utility of the data remains pending.
Written with AI from the reporting · updated 10/14/2026
Expand the detailed fact route and sources
Story route
A station-by-station route from the first public evidence onward; every station carries its own evidence and uncertainty.
- 1Unclassified fact stationOct 15, 00:00recorded Oct 15, 06:24single source
发布Minos预印本、博客和开放数据集
Read original source ↗ Latent SpaceReplay this moment in the explorerEvidence & revisions · 1 source records
- Subject：
- thinking-machines
- Action：
- 发布预印本、博客和开放数据集
- Object：
- AI playground for agents to do verifiable scientific reasoning
- Occurred (fact side)：
- Oct 15, 00:00
- Published (valid side)：
- Oct 15, 04:35
- First seen (known side)：
- Oct 15, 06:24
Fact stationSingle source · one-sided claimCopy may have been revised later; time replay is bounded. Layer labels group reading only, not causation between layers.
Uncertainty marks flag where to check the sources; they are not proven causation.
Report timeline
Explore every side through dated source reports.
- Latent SpaceRelated fact: single sourceLeash Bio releases Minos model and open dataset for lab-in-the-loop scientific reasoning
Report first recorded Oct 15, 06:24
Leash Bio, specializing in DNA Encoded Libraries (DELs) for high-throughput protein-small molecule interaction data, has released the Minos pre-trained model and an open dataset containing 6.5M molecules. This development enables 'lab-in-the-loop' simulations for scientific Reinforcement Learning (RL) agents, addressing previous industry challenges related to data quality and reproducibility.
Read original source ↗
Event discussion
Reader opinions are hosted on GitHub Discussions; posting requires a GitHub account. Comments are not verified facts.