Beyond Independent Genes: Learning Module-Inductive Representations for Gene Perturbation Prediction

TL;DR

This paper introduces scBIG, a novel framework that models coordinated gene programs to improve the prediction of transcriptional responses to genetic perturbations, especially in unseen and combinatorial cases.

Contribution

scBIG explicitly models gene programs and their interactions, overcoming limitations of gene-wise approaches and static priors, leading to better perturbation response predictions.

Findings

Outperforms state-of-the-art methods on multiple benchmarks.

Achieves 6.7% average improvement over strong baselines.

Effective on unseen and combinatorial perturbation scenarios.

Abstract

Predicting transcriptional responses to genetic perturbations is a central problem in functional genomics. In practice, perturbation responses are rarely gene-independent but instead manifest as coordinated, program-level transcriptional changes among functionally related genes. However, most existing methods do not explicitly model such coordination, due to gene-wise modeling paradigms and reliance on static biological priors that cannot capture dynamic program reorganization. To address these limitations, we propose scBIG, a module-inductive perturbation prediction framework that explicitly models coordinated gene programs. scBIG induces coherent gene programs from data via Gene-Relation Clustering, captures inter-program interactions through a Gene-Cluster-Aware Encoder, and preserves modular coordination using structure-aware alignment objectives. These structured representations are then modeled using conditional flow matching to enable flexible and generalizable perturbation prediction. Extensive experiments on multiple single-cell perturbation benchmarks show that scBIG consistently outperforms state-of-the-art methods, particularly on unseen and combinatorial perturbation settings, achieving an average improvement of 6.7% over the strongest baselines.