アブストラクト
| Date | 14:00-15:00 Aug 27, 2025 |
| Speaker | Xiaoqing Cheng School of Mathematics and Statistics, Xi'an Jiaotong University, China |
| Title | Identifying spatial domains under a contrastive learning framework |
| Abstract | Spatial transcriptomics is a rapidly growing field that aims to comprehensively characterize tissue organization and architecture at the single-cell or subcellular resolution using spatial information. Such techniques provide a solid foundation for the mechanistic understanding of many biological processes in both health and disease that cannot be obtained using traditional technologies. Several methods have been proposed to decipher the spatial context of spots in tissue using spatial information. In this talk, we would introduce two spatial domain identification models under the contrastive learning framework. One is ConSpaS, which is a novel node representation learning framework that precisely deciphers spatial domains by integrating local and global similarities based on graph autoencoder and contrastive learning. We propose an augmentation-free mechanism to construct global positive samples and use a semi-easy sampling strategy to define negative samples. Another is SpaConTDS, a multimodal contrastive learning method that fully utilize histology image by weighting image features based on the contribution of stained image information, and constructing negative samples through the tuple disturbed strategy to compensate for weak modalities (histology images). SpaConTDS also use reinforce algorithm to update the hyperparameters which balance the information in gene expression matrix and histology images, to achieve more accurate recognition in the spatial domain. We validated ConSpaS and SpaCOnTDS on multiple tissue types and technology platforms by comparing it with existing typical methods. The experimental results confirmed that they effectively improved the identification accuracy of spatial domains with biologically meaningful spatial patterns, and denoised gene expression data while maintaining the spatial expression pattern. |