MagNet: Multi-Level Attention Graph Network for Predicting High-Resolution Spatial Transcriptomics

Junchao Zhu, Ruining Deng, Tianyuan Yao, Juming Xiong, Chongyu Qu, Junlin Guo, Siqi Lu, Yucheng Tang, Daguang Xu, Mengmeng Yin, Yu Wang, Shilin Zhao, Yaohong Wang, Haichun Yang, Yuankai Huo
Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, PMLR 301:1943-1955, 2026.

Abstract

The rapid development of spatial transcriptomics (ST) offers new opportunities to explore the gene expression patterns within the spatial microenvironment. Current research integrates pathological images to infer gene expression, addressing the high costs and time-consuming processes to generate spatial transcriptomics data. However, as spatial transcriptomics resolution continues to improve, existing methods remain primarily focused on gene expression prediction at low-resolution (55$\mu$m) spot levels. These methods face significant challenges, especially the information bottleneck, when they are applied to high-resolution (8$\mu$m) HD data. To bridge this gap, this paper introduces MagNet, a multi-level attention graph network designed for accurate prediction of high-resolution HD data. MagNet employs cross-attention layers to integrate features from multi-resolution image patches hierarchically and utilizes a GAT-Transformer module to aggregate neighborhood information. By integrating multilevel features, MagNet overcomes the limitations posed by low-resolution inputs in predicting high-resolution gene expression. We systematically evaluated MagNet and existing ST prediction models on both a private spatial transcriptomics dataset and a public dataset at three different resolution levels. The results demonstrate that MagNet achieves state-of-the-art performance at both spot level and high-resolution bin levels, providing a novel methodology and benchmark for future research and applications in high-resolution HD-level spatial transcriptomics. Code is available at https://github.com/Junchao-Zhu/MagNet.

Cite this Paper

BibTeX
@InProceedings{pmlr-v301-zhu26b, title = {MagNet: Multi-Level Attention Graph Network for Predicting High-Resolution Spatial Transcriptomics}, author = {Zhu, Junchao and Deng, Ruining and Yao, Tianyuan and Xiong, Juming and Qu, Chongyu and Guo, Junlin and Lu, Siqi and Tang, Yucheng and Xu, Daguang and Yin, Mengmeng and Wang, Yu and Zhao, Shilin and Wang, Yaohong and Yang, Haichun and Huo, Yuankai}, booktitle = {Proceedings of The 8th International Conference on Medical Imaging with Deep Learning}, pages = {1943--1955}, year = {2026}, editor = {Tasdizen, Tolga and Elhabian, Shireen and Summers, Ronald and Chen, Chen and Koch, Lisa and Zhuang, Yan}, volume = {301}, series = {Proceedings of Machine Learning Research}, month = {09--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v301/main/assets/zhu26b/zhu26b.pdf}, url = {https://proceedings.mlr.press/v301/zhu26b.html}, abstract = {The rapid development of spatial transcriptomics (ST) offers new opportunities to explore the gene expression patterns within the spatial microenvironment. Current research integrates pathological images to infer gene expression, addressing the high costs and time-consuming processes to generate spatial transcriptomics data. However, as spatial transcriptomics resolution continues to improve, existing methods remain primarily focused on gene expression prediction at low-resolution (55$\mu$m) spot levels. These methods face significant challenges, especially the information bottleneck, when they are applied to high-resolution (8$\mu$m) HD data. To bridge this gap, this paper introduces MagNet, a multi-level attention graph network designed for accurate prediction of high-resolution HD data. MagNet employs cross-attention layers to integrate features from multi-resolution image patches hierarchically and utilizes a GAT-Transformer module to aggregate neighborhood information. By integrating multilevel features, MagNet overcomes the limitations posed by low-resolution inputs in predicting high-resolution gene expression. We systematically evaluated MagNet and existing ST prediction models on both a private spatial transcriptomics dataset and a public dataset at three different resolution levels. The results demonstrate that MagNet achieves state-of-the-art performance at both spot level and high-resolution bin levels, providing a novel methodology and benchmark for future research and applications in high-resolution HD-level spatial transcriptomics. Code is available at https://github.com/Junchao-Zhu/MagNet.} }
Endnote
%0 Conference Paper %T MagNet: Multi-Level Attention Graph Network for Predicting High-Resolution Spatial Transcriptomics %A Junchao Zhu %A Ruining Deng %A Tianyuan Yao %A Juming Xiong %A Chongyu Qu %A Junlin Guo %A Siqi Lu %A Yucheng Tang %A Daguang Xu %A Mengmeng Yin %A Yu Wang %A Shilin Zhao %A Yaohong Wang %A Haichun Yang %A Yuankai Huo %B Proceedings of The 8th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Tolga Tasdizen %E Shireen Elhabian %E Ronald Summers %E Chen Chen %E Lisa Koch %E Yan Zhuang %F pmlr-v301-zhu26b %I PMLR %P 1943--1955 %U https://proceedings.mlr.press/v301/zhu26b.html %V 301 %X The rapid development of spatial transcriptomics (ST) offers new opportunities to explore the gene expression patterns within the spatial microenvironment. Current research integrates pathological images to infer gene expression, addressing the high costs and time-consuming processes to generate spatial transcriptomics data. However, as spatial transcriptomics resolution continues to improve, existing methods remain primarily focused on gene expression prediction at low-resolution (55$\mu$m) spot levels. These methods face significant challenges, especially the information bottleneck, when they are applied to high-resolution (8$\mu$m) HD data. To bridge this gap, this paper introduces MagNet, a multi-level attention graph network designed for accurate prediction of high-resolution HD data. MagNet employs cross-attention layers to integrate features from multi-resolution image patches hierarchically and utilizes a GAT-Transformer module to aggregate neighborhood information. By integrating multilevel features, MagNet overcomes the limitations posed by low-resolution inputs in predicting high-resolution gene expression. We systematically evaluated MagNet and existing ST prediction models on both a private spatial transcriptomics dataset and a public dataset at three different resolution levels. The results demonstrate that MagNet achieves state-of-the-art performance at both spot level and high-resolution bin levels, providing a novel methodology and benchmark for future research and applications in high-resolution HD-level spatial transcriptomics. Code is available at https://github.com/Junchao-Zhu/MagNet.
APA
Zhu, J., Deng, R., Yao, T., Xiong, J., Qu, C., Guo, J., Lu, S., Tang, Y., Xu, D., Yin, M., Wang, Y., Zhao, S., Wang, Y., Yang, H. & Huo, Y.. (2026). MagNet: Multi-Level Attention Graph Network for Predicting High-Resolution Spatial Transcriptomics. Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 301:1943-1955 Available from https://proceedings.mlr.press/v301/zhu26b.html.

Related Material