Welcome to stCluster’s documentation!
Overview
Spatial transcriptomics is a powerful technique that provides insights into gene expression in the context of native tissues. It bridges molecular data with spatial information, uncovering intricate cellular interactions and tissue organization. To decode cellular spatial domains, integration of gene expression data and spatial information is essential. stCluster is introduced as a novel method that enhances the identification of spatial domains using graph contrastive learning and multi-task learning. It refines informative representations for spatial transcriptomic data, enabling the recognition of spatially coherent patterns. Through optimization of multiple tasks, stCluster can capture complex relationships between gene expression and spatial organization. Experimental results demonstrate its proficiency in accurately identifying complex spatial domains across various datasets and platforms, including tissues, organs, and embryos, surpassing existing state-of-the-art methods. Additionally, stCluster is capable of denoising spatial gene expression patterns and enhancing spatial trajectory inference.
Note
To fully reproduce the results as described in the paper, it is recommended to use the container we have provided on a Nvidia RTX 3090 GPU device