SpatialMETA
Integrating Cross-Sample and Cross-Modality Data for Spatial Transcriptomics and Metabolomics with CVAE
SpatialMETA: Integrating Cross-Sample and Cross-Modal Data for Spatial Transcriptomics and Metabolomics
Status: Featured co-author publication in Nature Communications (2025)
Role: Research Assistant, Co-author
Institution: ZJU-UoE Institute, Zhejiang University; collaboration with Prof. Wanlu Liu’s laboratory
Duration: November 2024 – Spring 2025
Project Overview
SpatialMETA is a computational framework for integrating spatial transcriptomics and spatial metabolomics across samples and modalities. The core method uses a conditional variational autoencoder to align heterogeneous spatial signals into a shared analytical space, enabling more coherent downstream biological interpretation.
My Contributions
- Benchmark design: built benchmark settings and evaluation criteria for cross-sample, cross-modal integration
- Method validation: assessed robustness and biological plausibility across multiple experimental settings
- Quantitative analysis: designed metrics to compare integration quality and downstream utility
Technical Approach
The method focuses on:
- conditional variational autoencoder-based multimodal alignment
- cross-sample integration of spatial transcriptomics and metabolomics
- quantitative evaluation of integration quality for downstream spatial analysis
Outcome
This project became one of my featured co-author publications and strengthened my interest in building rigorous computational frameworks for spatial multi-omics analysis.
Publication: Tian R†, Xue Z†, Chen Y, Qi Y, Zhang J, Yuan J, Ruan D, Lin J, Liu J, Wang D, Youqiong Y, Liu W*. Integrating Cross-Sample and Cross-Modal Data for Spatial Transcriptomics and Metabolomics with SpatialMETA. *Nature Communications*, 2025.