Across several benchmarks, this method outperforms the strongest published method in settings involving combinatorial and unseen perturbation prediction, and is the strongest published method in settings involving combinatorial and unseen perturbation prediction.
Abstract
Predicting cellular responses to perturbations is a central problem in cellular biology, with broad applications in systems biology and drug discovery. This task is challenging because cellular responses can be complex and cell-state dependent, intrinsic cell-to-cell variability can be confounded with perturbation effects, and destructive single-cell RNA sequencing precludes paired measurements of the same cell before and after treatment. Flow matching transports control cells to perturbed states flexibly, but acting on the full cell state can confound perturbation effects with pre-existing cell-to-cell variability. Disentangled approaches separate responsive from invariant components, but model perturbations through prescribed mechanisms, such as latent shifts or graph edits, limiting their flexibility. We address both limitations in a unified framework. A variational encoder disentangles each cell into an invariant block, capturing state unaffected by the perturbation, and a responsive block, capturing state it changes, through conditional priors and an information-theoretic invariance constraint. Conditional flow matching transports only the responsive block, conditioned on the perturbation and invariant state, yielding a flexible, data-driven model of perturbation effects without confounding pre-existing variability. Across several benchmarks, our method outperforms the strongest published method in settings involving combinatorial and unseen perturbation prediction.
PerturbBridge is proposed, a conditional latent Schrödinger Bridge framework that reformulates stochastic population transport over high-dimensional, sparse geneexpression profiles as bridge learning in a compact cell latent space, and achieves state-of-the-art performance in differential-expression recovery on both be...
Zhi-Hao Liu, Chang-Zhi Jiang, Can Yang et al.· bioRxiv· 0 citations
This work introduces a diffusion-based generative model that learns explicit gene-to-gene regulation directly from perturbation data and improves the prediction of combinatorial perturbation effects and identifies causal links that standard statistical comparisons miss.
Lars Lorch, Jia-Qi Zhang, Charlotte Bunne et al.· Proceedings of the National...· 9 citations· ⚡2
Overall, scLDM provides a robust and biologically consistent strategy for in silico perturbation screening, and exhibits strong interpretability, as the learned perturbation embeddings show high functional alignment with known biological mechanisms.
PerturbLDM, a latent-diffusion framework for conditional generation of single-cell transcriptional responses, is introduced, showing support for conditional response generation across data scales and biological settings.
Li-Shan Yu, Kang-Lin Hsieh, Y. Chu et al.· bioRxiv· 0 citations
Predicting phenotypic and transcriptional responses to perturbations at single-cell resolution provides a powerful tool for probing biological systems. However, existing methods typically rely on fixed mappings learned during training, making it challenging to calibrate distribution shifts or adapt to novel perturbatio...
Jia-Nan Wei, Jia-Jun Hong, Gui-Kun Chen et al.· 0 citations
A mechanism-centric virtual cell world model that represents cellular state as a set of Latent Mechanism Units (LMUs) and treats genetic perturbations as actions on these latent states supports mechanism-level latent state transition as a useful formulation for virtual cell models that aim to predict and interpret cell...
Yu-Wei Miao, Azim Dehghani Amirabad, S. Oloff et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026