Multi-scale microenvironment covariance for unsupervised spatial domain identification
- 1Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, 277-8563, C
Abstract
Motivation: Many spatial domain identification methods describe the neighbourhood of a cell by mean gene expression. However, the mean cannot separate neighbourhoods that differ only in gene-gene covariance and tissue is organised at several spatial scales. Results: We propose CoSMIM, which represents the niche of each cell by a weighted fusion of gene-gene covariance matrices of its spatial neighbourhood at multiple scales. Principal component analysis and Leiden clustering of the matrix square root of the fused matrix identify spatial domains without model training. On six benchmark datasets from five platforms, CoSMIM attains the best overall rank over eight evaluation metrics among 16 methods. It also ranks first on a Xenium mouse brain dataset, and on all seven datasets its median NMI exceeds that of single-scale COVET. The domain covariance matrices also show gene pairs whose correlation differs between domains.
Related
- HESTIA: Scalable Multimodal Integration of Histology and High-Resolution Spatial Transcriptomics for Robust Spatial Domain Identification
- Kontext: Kernel-based method bridging spatial domain identification and cell-cell communication
- SpaMOAL is a deep learning method that enables accurate spatial domain identification from multi-omics data
- Simple but effective unsupervised classification for specified domain images: a case study on fungi images
- FFPERescuer: deep unsupervised domain adaptation for the reconstruction of gene expression profiles derived from formalin-fixed paraffin-embedded samples
BibTeX
@misc{du2026multiscale,
title = {{Multi-scale microenvironment covariance for unsupervised spatial domain identification}},
author = {B. Du and H. Kiryu},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.20.760373},
url = {https://www.biorxiv.org/content/10.64898/2026.10.20.760373v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Multi-scale microenvironment covariance for unsupervised spatial domain identification
AU - B. Du
AU - H. Kiryu
PY - 2026
DA - 2026/10/20
PB - bioRxiv
DO - 10.64898/2026.10.20.760373
UR - https://www.biorxiv.org/content/10.64898/2026.10.20.760373v1
AB - Motivation: Many spatial domain identification methods describe the neighbourhood of a cell by mean gene expression. However, the mean cannot separate neighbourhoods that differ only in gene-gene covariance and tissue is organised at several spatial scales. Results: We propose CoSMIM, which represents the niche of each cell by a weighted fusion of gene-gene covariance matrices of its spatial neighbourhood at multiple scales. Principal component analysis and Leiden clustering of the matrix square root of the fused matrix identify spatial domains without model training. On six benchmark datasets from five platforms, CoSMIM attains the best overall rank over eight evaluation metrics among 16 methods. It also ranks first on a Xenium mouse brain dataset, and on all seven datasets its median NMI exceeds that of single-scale COVET. The domain covariance matrices also show gene pairs whose correlation differs between domains.
ER -