StereoCell enables high accuracy single cell segmentation for spatial transcriptomic dataset

With recent advances in resolution and field-of-view, spatially resolved sequencing has emerged as a cutting-edge technology that provides a technical foundation for interpreting large tissues at the spatial single-cell level. To handle the high-resolution spatial omics dataset with associated images and generate spatial single-cell level gene expression, a powerful one-stop toolbox is required. Here, we propose StereoCell, an image-facilitated cell segmentation framework for high-resolution and large field-of-view spatial omics. StereoCell offers a comprehensive and systematic solution to generating high-confidence spatial single-cell data, including image stitching, registration, nuclei segmentation, and molecule labeling. In image stitching and molecule labeling, StereoCell delivers the best-performing algorithms to reduce stitching error and improve the signal-to-noise ratio of single-cell gene expression compared to existing methods. Meanwhile, as demonstrated using mouse brain, StereoCell has been shown to obtain high-accuracy spatial single-cell data, which facilitates clustering and annotation.