STOmics technology:BGI Stereomics Stereo-Seq

Data type:Spatial transcriptomics

Sample scope:Other

Summary:Spatial transcriptomics (ST) offers the opportunity to study gene expression patterns and the spatial microenvironment in situ. However, current ST sequencing technologies are often limited by low detection sensitivity and restricted gene throughput. Achieving precise cell type annotations for each cell within spatial transcriptomic data, which is essential for understanding biological processes at the single-cell level, remains a significant challenge. In this work, we present stTransfer, a novel method for annotating single cells in spatial transcriptomic data that integrates graph neural networks and transfer learning. This approach leverages existing information from reference single-cell RNA sequencing (scRNA-seq) datasets as well as the spatial context provided by spatial transcriptomics. Through a series of benchmark analyses on publicly available spatial transcriptomic datasets, we demonstrate that stTransfer outperforms other state-of-the-art methods in terms of accuracy and robustness. Furthermore, we applied stTransfer to a self-collected dataset of the zebra finch optic tectum, obtained using Stereo-seq technology. Our analysis led to the discovery of a distinct class of neurons, highlighting the potential of our method to uncover new insights into cellular diversity and organization.

Contributor(s):tao zhou.

Submitter:zhoutao(zhoutao),bgi

Release date:2025-01-31

Updated:2025-10-10