In-silico perturbation search over the cellular state space
L. Chen · Bio Group
Biuron Technical Notes
Given a learned model of cellular dynamics, discovery becomes search: find the perturbation that drives the cell toward a target state. We formalise this as optimisation over the model's latent space and validate the proposals against held-out experiments.
The method recovers established interventions and proposes novel ones, several of which reproduced in follow-up assays.
This is a plain-language summary. The full manuscript and supplementary material are available on request.