SPACE: Spatially variable gene clustering adjusting for cell type effect for improved spatial domain detection.

Recent advances in spatial transcriptomics have significantly deepened our understanding of biology. A primary focus has been identifying spatially variable genes (SVGs) which are crucial for downstream tasks like spatial domain detection. Traditional methods often use all or a set number of top SVGs for this purpose. However, in diverse datasets with many SVGs, this approach may not ensure accurate results. Instead, grouping SVGs by expression patterns and using all SVG groups in downstream analysis can improve accuracy. Furthermore, classifying SVGs in this manner is akin to identifying cell type marker genes, offering valuable biological insights. The challenge lies in accurately categorizing SVGs into relevant clusters, aggravated by the absence of prior knowledge regarding the number and spectrum of spatial gene patterns. Addressing this challenge, we propose SPACE, SPatially variable gene clustering Adjusting for Cell type Effect, a framework that classifies SVGs based on their spatial patterns by adjusting for confounding effects caused by shared cell types, to improve spatial domain detection. This method does not require prior knowledge of gene cluster numbers, spatial patterns, or cell type information. Our comprehensive simulations and real data analyses demonstrate that SPACE is an efficient and promising tool for spatial transcriptomics analysis. Key PointsO_LISPACE eliminates the need for prior knowledge about the number of gene clusters, known cell types, or the quantity of SVGs to identify clusters for downstream analysis. C_LIO_LISPACE offers a method to effectively leverage SVGs for low-dimensional embedding within each cluster to improve the accuracy of spatial domain detection. C_LIO_LIThe efficiency and utility of the SPACE algorithm have been validated across multiple datasets and simulations, demonstrating its effectiveness in producing meaningful and interpretable results. C_LI