SpaFactor: Lightweight Spatial Context-Aware Gene Program Modeling for Histology-to-Transcriptomics Inference
Abstract
Spatial transcriptomics (ST) profiles gene expression within tissue architecture, but its cost and experimental complexity limit routine use. Predicting spatial expression from routinely available hematoxylin and eosin (HE) images therefore offers a scalable alternative. However, conventional methods often fit high-dimensional gene outputs as independent targets, overlooking the biological coordination among genes while remaining vulnerable to high-dimensional noise and overfitting. Existing attempts to address this limitation often rely on computationally heavy graph networks or complex auxiliary supervision. We therefore introduce SpaFactor, a lightweight and efficient low-rank morphology-program-gene factorization framework. At the input, SpaFactor efficiently fuses the visual representation of the central spot with multiscale local and regional neighborhood context, yielding a histologic representation that captures cellular morphology and microenvironmental heterogeneity. For modeling, a residual MLP stably learns a nonlinear mapping from the tissue microenvironment to low-dimensional latent gene programs. These activities are decoded through shared gene loadings into coordinated multi-gene expression predictions. Across five public cohorts, SpaFactor achieves the best aggregate performance, with particularly clear improvements for spatially variable genes, and more faithfully recovers biologically organized spatial patterns. These results demonstrate that lightweight joint modeling of tissue context and gene programs can improve both predictive accuracy and biological fidelity.
Related
- HESTIA: Scalable Multimodal Integration of Histology and High-Resolution Spatial Transcriptomics for Robust Spatial Domain Identification
- Probabilistic uncertainty in virtual spatial transcriptomics from histology is largely redundant with the predicted mean
- VGL-stGen: Learning a shared multimodal model for lung histology and spatial transcriptomics
- MorphoProt: RNA-Conditioned Integration of Multiscale Histology for Spatial Protein Inference
- Quantitative Donor Histology and Transcriptomics Link Pancreatic Composition and Precancers Across the Adult Lifespan
BibTeX
@misc{ruan2026spafactor,
title = {{SpaFactor: Lightweight Spatial Context-Aware Gene Program Modeling for Histology-to-Transcriptomics Inference}},
author = {Shiting Ruan and Xitong Ling and Qiming He and Ziyou Yan and Huaitian Yuan and Tian Guan and Ying Xiao and Xu Guan and Yonghong He},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.28563},
eprint = {2609.28563},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.28563},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - SpaFactor: Lightweight Spatial Context-Aware Gene Program Modeling for Histology-to-Transcriptomics Inference
AU - Shiting Ruan
AU - Xitong Ling
AU - Qiming He
AU - Ziyou Yan
AU - Huaitian Yuan
AU - Tian Guan
AU - Ying Xiao
AU - Xu Guan
AU - Yonghong He
PY - 2026
DA - 2026/09/23
PB - arXiv
DO - 10.48550/arxiv.2609.28563
UR - https://arxiv.org/abs/2609.28563
AB - Spatial transcriptomics (ST) profiles gene expression within tissue architecture, but its cost and experimental complexity limit routine use. Predicting spatial expression from routinely available hematoxylin and eosin (HE) images therefore offers a scalable alternative. However, conventional methods often fit high-dimensional gene outputs as independent targets, overlooking the biological coordination among genes while remaining vulnerable to high-dimensional noise and overfitting. Existing attempts to address this limitation often rely on computationally heavy graph networks or complex auxiliary supervision. We therefore introduce SpaFactor, a lightweight and efficient low-rank morphology-program-gene factorization framework. At the input, SpaFactor efficiently fuses the visual representation of the central spot with multiscale local and regional neighborhood context, yielding a histologic representation that captures cellular morphology and microenvironmental heterogeneity. For modeling, a residual MLP stably learns a nonlinear mapping from the tissue microenvironment to low-dimensional latent gene programs. These activities are decoded through shared gene loadings into coordinated multi-gene expression predictions. Across five public cohorts, SpaFactor achieves the best aggregate performance, with particularly clear improvements for spatially variable genes, and more faithfully recovers biologically organized spatial patterns. These results demonstrate that lightweight joint modeling of tissue context and gene programs can improve both predictive accuracy and biological fidelity.
ER -