Academic Report of School of Mathematical Sciences [2026] No. 033

(Series Report for High-Level University Construction No. 1292)

Title：Deciphering Nuclei-Resolved Spatial Biology by Uniting Spatial Omics and Histopathology

Speaker：Lingyu Li (The University of Hong Kong)

Time：10:30-11:30, Apr. 26, 2026

Location：Huixing Building 514

Abstract：High-resolution spatial transcriptomics (ST) and multiplexed imaging are transforming our understanding of the tumor microenvironment, but their cost, data sparsity, and computational burden still limit routine use. In this talk, I will present two complementary computational frameworks, FineST and SparseAEH, that together provide a scalable pipeline from standard histopathology to nuclei-resolved maps of cell communication, phenotypes, and gene expression programs. FineST uses deep contrastive learning to fuse ST with matched histology via histology foundation models, enabling accurate nuclei segmentation, high-resolution RNA imputation, and ligand-receptor analysis at single-nucleus resolution. Across VisiumHD and Xenium datasets in multiple cancer types, FineST consistently improves RNA imputation, cell-type prediction, and the discovery of cell-cell communication. Complementing this, SparseAEH addresses the computational bottleneck in discovering spatial expression “histologies” from dense ST data by using a Gaussian process mixture model with block-wise sparse covariance, enabling scalable spatial gene clustering with substantial speedups while preserving accuracy. Together, these components form a coherent system that integrates histology, spatial omics, and scalable probabilistic modeling to deliver nuclei-resolved, interpretable spatial maps of cell communication and expression patterns for biological discovery and clinical translation.This talk is based on joint work with Tianjie Wang, under the supervision of Prof. Yuanhua Huang.

Speaker Profile: Lingyu Li is a postdoctoral researcher at the Faculty of Medicine, The University of Hong Kong, under the supervision of Associate Professor Yuanhua Huang. She received her Ph.D. from the Department of Biomedical Engineering at Shandong University in 2023. During her studies, she was supported by an overseas exchange scholarship from 2021 to 2023 to pursue joint training at the Department of Mathematics, The University of Hong Kong. Her research focuses on bioinformatics, specifically the development of biostatistical models and machine learning algorithms, with a commitment to addressing key scientific challenges such as biomarker discovery, the inference of gene regulatory networks, and the identification of drug resistance-driving genes. During his postdoctoral research, he has primarily focused on multimodal artificial intelligence, spatial transcriptomics, and medical pathology image analysis. He has published over ten papers as first author in journals such as Nature Communications , npj Digital Medicine , and Bioinformatics , as well as four co-authored papers, with a total citation count exceeding 360. He delivered a 20-minute long talk at the ISMB 2024 international conference and received a Travel Fellowship from the International Society for Computational Biology (ISCB). He is currently a member of the International Society for Computational Biology and the Chinese Society for Industrial and Applied Mathematics, and serves on the Program Committee for IEEE BIBM 2026; he has also been invited to serve as a reviewer for several renowned journals, including Cell , Genome Research , and Bioinformatics, as well as for relevant CCF conferences.

Faculty and students are welcome to attend!

Invited by: Yushan Qiu

School of Mathematical Sciences

April 22, 2026