Causality-integrated graph learning for multi-endpoint toxicity prediction
Proceedings of the National Academy of Sciences, Volume 123, Issue 41, October 2026. SignificanceExisting models largely predict toxicity one end point at a time, although chemical toxicity emerges from coordinated perturbations across multiple targets, pathways, and adverse outcomes. We developed a causality-integrated graph learning ...
We haven't written up this one. PNAS has the full story — the link below goes straight to it.
