ORBIT: Learning Gene Program Co-Activation Structure for Cell-Type-Stratified Pathway Rewiring Analysis in Single-Cell Transcriptomics
Yuechen Wang, Lina Jia, Qinglong Wang, Feng Tian
TLDR
ORBIT is a self-supervised transformer that learns asymmetric gene program dependencies from single-cell RNA-seq, revealing cell-type-specific pathway rewiring.
Key contributions
- Learns asymmetric gene program dependencies from observational scRNA-seq using a self-supervised transformer.
- Employs an intervention-consistent objective to model directional influence, not just symmetric co-occurrence.
- Recovers co-activation structures consistent with established Alzheimer's disease vulnerability signatures.
- Identifies cell-type-specific pathway rewiring invisible to standard differential expression methods.
Why it matters
This paper introduces a novel self-supervised approach to model complex, asymmetric gene program interactions, a significant advance over methods treating programs independently or requiring perturbations. It provides deeper insights into cell-type-specific pathway rewiring in diseases like Alzheimer's, improving our understanding of disease mechanisms.
Original Abstract
Gene programs co-activate within cells, but existing single-cell methods either treat programs independently or require experimental perturbation data to model their interactions. We introduce ORBIT, a self-supervised transformer that learns asymmetric dependencies among gene programs from observational single-cell RNA-sequencing data alone, quantifying how strongly each program influences every other program. The key mechanism is an intervention-consistent training objective: the model learns each program's directional influence on every other program by predicting how the others change when that program is removed, yielding attention weights that reflect asymmetric influence rather than symmetric co-occurrence. Applied to 191,890 prefrontal cortex nuclei across three pathway vocabularies, ORBIT recovers co-activation structure consistent with established Alzheimer's disease vulnerability signatures, identifies cell-type-specific rewiring invisible to differential expression, and achieves 0.984 macro F1 on cell-type classification from 220 pathway scores, which is within 0.3 points of a state-of-the-art classifier using all 22,088 genes.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.