NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics
- 1The University of Manchester
Abstract
Spatial transcriptomics captures molecular states within cells and their organisation in tissue. However, integrating fine-grained gene information with spatial context at scale remains challenging for existing foundation models. Here we present NexuST, a hierarchical foundation model that repeatedly interleaves gene-level molecular modelling with cell-level spatial modelling, allowing the two levels to refine one another during end-to-end pretraining. For pretraining, we curated HumanST-46M, comprising 45.7 million human cells from 72 datasets across 11 organs and three imaging-based platforms. Across four held-out datasets totalling approximately 2.6 million cells, NexuST achieved state-of-the-art or competitive performance in cell-type annotation, region prediction, gene recovery and neighbourhood-composition prediction. We find that cell-intrinsic expression remains informative even for spatial tasks, as shown by an expression-only PCA baseline, while NexuST shows particularly strong gains where spatial context is essential. Overall, NexuST establishes a hierarchical framework that can serve as a general backbone for future spatial transcriptomics foundation models.
Related
- Harmonised benchmarking of foundation models for single-cell and spatial transcriptomics reveals context-dependent generalisation
- stEDGE enables edge-guided multiscale reconstruction of hierarchical spatial domains and transition interfaces in spatial transcriptomics
- Spatial transcriptomics profiling across tissue boundaries in the developing mouse brain
- SpaCoEx: Sparse Gene Selection for Spatially Varying Co-expression in Spatial Transcriptomics
- NicheCorr: Identifying Niche-Specific Differential Gene Co-Expression in Spatial Transcriptomics
BibTeX
@misc{liu2026nexust,
title = {{NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics}},
author = {H. Liu and Q. Zhao and L. Lin and Z. Zou and W. Cai and J. Sun and Y. Zhou and M. A. Alvarez and A. Gilmore and M. Rattray and A. F. Frangi and H. Zhou},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.22.753590},
url = {https://www.biorxiv.org/content/10.64898/2026.09.22.753590v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics
AU - H. Liu
AU - Q. Zhao
AU - L. Lin
AU - Z. Zou
AU - W. Cai
AU - J. Sun
AU - Y. Zhou
AU - M. A. Alvarez
AU - A. Gilmore
AU - M. Rattray
AU - A. F. Frangi
AU - H. Zhou
PY - 2026
DA - 2026/09/22
PB - bioRxiv
DO - 10.64898/2026.09.22.753590
UR - https://www.biorxiv.org/content/10.64898/2026.09.22.753590v1
AB - Spatial transcriptomics captures molecular states within cells and their organisation in tissue. However, integrating fine-grained gene information with spatial context at scale remains challenging for existing foundation models. Here we present NexuST, a hierarchical foundation model that repeatedly interleaves gene-level molecular modelling with cell-level spatial modelling, allowing the two levels to refine one another during end-to-end pretraining. For pretraining, we curated HumanST-46M, comprising 45.7 million human cells from 72 datasets across 11 organs and three imaging-based platforms. Across four held-out datasets totalling approximately 2.6 million cells, NexuST achieved state-of-the-art or competitive performance in cell-type annotation, region prediction, gene recovery and neighbourhood-composition prediction. We find that cell-intrinsic expression remains informative even for spatial tasks, as shown by an expression-only PCA baseline, while NexuST shows particularly strong gains where spatial context is essential. Overall, NexuST establishes a hierarchical framework that can serve as a general backbone for future spatial transcriptomics foundation models.
ER -