AnnFlux: object-conditioned neural stochastic differential equations for single-cell perturbation dynamics
- 1Department of Integrated Biomedical and Life Science, Korea University, Seoul, 02841, Republic of Korea
Abstract
Single-cell perturbation profiling measures responses to genetic and chemical interventions, yet most models learn a static map, ignoring how populations move over time and how perturbations combine. AnnFlux, an object-conditioned stochastic differential equation, learns a drift field in latent cell-state space. Conditioning on the perturbing object makes the field queryable one object at a time, yielding per-object drifts comparable across genes and drugs. By learning a drift field tailored to each perturbation context, it interpolates a held-out timepoint in an epithelial-mesenchymal transition time course and predicts unseen perturbations. Beyond point estimates, AnnFlux improves distributional fidelity and predicts responses to held-out perturbation combinations. An IFN-response signature predicted by AnnFlux was associated with TLS proximity in an independent pan-cancer spatial atlas. This framework maps perturbation-driven cell-state evolution as continuous trajectories and represents unseen perturbations using prior-knowledge embeddings.
Related
- A rank-conditioned α-stable perturbation for differential evolution: Separating perturbation frequency and tail heaviness
- GeoSinkhorn Flow: Geometry-Aware Flow Matching for Perturbation-Conditioned Dynamics in Single cell Data Phenoscapes
- Stochastic partial differential equation model for environmental DNA dynamics in river environments
- Chaotic dynamics in blooms of Ostreopsis cf. ovata modelled from environmental time series using ordinary differential equations systems
- Post-transcriptional Regulation of Stochastic Gene Expression Conditioned on Large Deviations
BibTeX
@misc{choi2026annflux,
title = {{AnnFlux: object-conditioned neural stochastic differential equations for single-cell perturbation dynamics}},
author = {H. Choi and G. Byeon and H. Park and J. Park and S. Lim and J.-Y. An},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.01.748703},
url = {https://www.biorxiv.org/content/10.64898/2026.09.01.748703v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - AnnFlux: object-conditioned neural stochastic differential equations for single-cell perturbation dynamics
AU - H. Choi
AU - G. Byeon
AU - H. Park
AU - J. Park
AU - S. Lim
AU - J.-Y. An
PY - 2026
DA - 2026/09/01
PB - bioRxiv
DO - 10.64898/2026.09.01.748703
UR - https://www.biorxiv.org/content/10.64898/2026.09.01.748703v1
AB - Single-cell perturbation profiling measures responses to genetic and chemical interventions, yet most models learn a static map, ignoring how populations move over time and how perturbations combine. AnnFlux, an object-conditioned stochastic differential equation, learns a drift field in latent cell-state space. Conditioning on the perturbing object makes the field queryable one object at a time, yielding per-object drifts comparable across genes and drugs. By learning a drift field tailored to each perturbation context, it interpolates a held-out timepoint in an epithelial-mesenchymal transition time course and predicts unseen perturbations. Beyond point estimates, AnnFlux improves distributional fidelity and predicts responses to held-out perturbation combinations. An IFN-response signature predicted by AnnFlux was associated with TLS proximity in an independent pan-cancer spatial atlas. This framework maps perturbation-driven cell-state evolution as continuous trajectories and represents unseen perturbations using prior-knowledge embeddings.
ER -