数苑经纬讲坛（112）：Integrative analysis and regulatory inference for spatial multi-omics and single-cell omics data
来源：
发布时间：2026-09-17
报告时间：- 2026年9月22日（周二）下午17:00-18:00
- 报告地点：
- 6号楼C203
- 报告人：
- 林志翔
- 报告人单位：
- 香港中文大学
- 邀请人：
- 张晓飞
- 摘要：
- In this presentation, I will focus on two of our recent works, MultiGATE and TIVelo. First, I will talk about MultiGATE, which is developed for analyzing spatial multi-omics data. New spatial multi-omics technologies, which jointly profiles transcriptomic and epigenome/proteome markers for the same tissue section, have expanded the frontiers of spatial techniques. MultiGATE utilizes a two-level graph attention auto-encoder to integrate the multi-modality and spatial information in spatial multi-omics data. The key feature of MultiGATE is that it simultaneously performs embedding of the spatial pixels and infers the cross-modality regulatory relationship, which allows deeper data integration and provides insights into transcriptional regulation. Through effectively integrating spatial multi-omics data, MultiGATE both enhances the extraction of latent embeddings of the pixels and boosts the inference of transcriptional regulation for cross-modality genomic features. Second, I will talk about TIVelo, which is developed for RNA velocity estimation. RNA velocity analysis has become an essential approach for characterizing dynamic cellular processes such as development, differentiation, and disease progression. However, most existing RNA velocity methods are constrained by explicit assumptions of ordinary differential equations (ODEs), limiting their ability to model complex transcriptomic expression patterns. Unlike previous methods, TIVelo employs an orientation score to determine cluster-wise directionality without relying on explicit ODE formulations, thereby accommodating more complex and realistic transcriptional behaviors. We demonstrate the effectiveness of TIVelo across 16 real-world datasets and benchmark its performance against six established methods, highlighting its improved capability to capture intricate gene expression dynamics that deviate from standard ODE-based assumptions.
林志翔（香港中文大学）
Zhixiang Lin is currently an associate professor in the Department of Statistics and Data Science at the Chinese University of Hong Kong. He received his B.S. from Tsinghua University in 2010 and Ph.D. from Yale University in 2015. Before joining CUHK, he was a postdoctoral researcher at Stanford University. His research focuses on developing novel statistical methods and computational tools for the analysis and interpretation of large-scale genomic data in biomedicine. More recently, his research group is working on AI for science.
