Getting Started

What is StereoMap?

StereoMap is a programming-free desktop application that allows you to interactively visualize and explore your Stereo-seq data from STOmics Products. You can use StereoMap to pinpoint the precise locations of genes within a tissue with subcellular resolution, discover molecular expression patterns for single or multiple genes in the original tissue section, and gain insights into the underlying biology.

A typical Stereo-seq data analysis workflow involves StereoMap at both upstream and downstream steps.

In the upstream, StereoMap offers Image QC and Image Processing tools to assess whether the image can be automatically processed by SAW or manually processed before analyzing with a spatial gene expression matrix. In the downstream, StereoMap provides Visual Explore functions for interactively visualizing and investigating data.

• Visual Explore: a key module of StereoMap for exploring Stereo-seq dataset. It enables interactive investigation of feature expression distribution across tissue within bins or cells. It also offers the ability to integrate images or multiple omics data for co-visualization or side-by-side visualization to inform downstream analysis.

• Image Processing: a core function for manipulating images. It allows for the manual alignment of an image with the feature expression matrix, and the execution of tissue and cell segmentation either manually or through the importation of result from external tools. The result can subsequently be transferred to the SAW analysis workflow for standard co-analysis.

• Tools: complementary tools support the processing of images and Stereo-seq datasets.

  • Stereo-seq Image QC: assesses the quality of microscope images and predicts their suitability for automated analysis in SAW.

---

StereoMap Input Files

The microscope image and SAW output files listed below are needed for image processing and visualization.

Visual Explore Input Files

Download and unzip the SAW output bundled visualization file (visualization.tar.gz), you will find a .stereo manifest file. The .stereo manifest file is required to open any visualization data in StereoMap. It contains information about where to find the visualization files packaged in the SAW output visualization.tar.gz file. To properly open data by manifest file, please follow the rules:

• Highly recommend keeping the default file structure of the SAW output directory. All the files are managed in the same directory path as the manifest file by default. If a file has been moved to another place or the file name has been changed, remember to modify the corresponding file path in the manifest file.

• At least a feature expression matrix file (.gef) or an image pyramid file (.rpi) must be available in the directory for the Visual Explore module.

File extension	Description
.stereo	A manifest file in JSON format that includes experiment and pipeline information, basic analysis statistics, and references to image and spatial matrix files in the SAW output visualization file folder. Unzip the SAW-generated bundle file (visualization.tar.gz) to find this .stereo manifest file. Required to open any SAW pipeline output in StereoMap.
.gef	The feature expression matrix file in HDF5 format for visualization. It contains the MID count for each gene of each spot. A spot is a binning unit that has a fixed-sized square shape in which the expression value in this square is accumulated. By default, a visualization .gef includes spot sizes of bin 1, 5, 10, 20, 50, 100, 200.
.cellbin.gef	The cell-feature expression matrix file in HDF5 format for visualization. It contains the spatial location and area of each cell, the MID count for each gene of each cell, and the cluster the cell belongs to. In .cellbin.gef, the cell is the smallest data unit. Only available when the image was processed in SAW pipelines.
.rpi	.rpi file saves one or multiple images in pyramidal format for better visualization. Each image is downsampled into several resolutions and each resolution layer is chopped into 256 pixels × 256 pixels tiles. If the size of a tile or a layer is smaller than 256 pixels × 256 pixels, it will remain intact. A typical SAW-generated .rpi organizes images and tiles sequentially by staining type (ssDNA, DAPI, H&E, or IF name) -> image type (registered image, tissue segmentation binary mask, or cell segmentation binary mask) -> resolution (equivalent resolution size of bin 2, bin 10, bin 50, and bin 100). Required to open any image in StereoMap Visual Explore.
<SN>.bin<N>_<leiden_res>.h5ad	A file stores clustering information from the spatial transcriptomics dataset in the AnnData file format. A <SN>.bin<N>_<leiden_res>.h5ad file is only allowed to contain the analysis results in one bin size. In the file name, <SN> stands for the Stereo-seq chip serial number, <N> for bin size, and <leiden_res> for leiden resolution. In a typical SAW pipeline, the spatial clustering analysis is processed with a spatial coordinate bin size of 200 and the Leiden resolution of 1.0.
<SN>.cellbin_<leiden_res>*.h5ad	The result of cell clustering information in AnnData file format. In the file name, <SN> represents the Stereo-seq chip serial number, <leiden_res> corresponds to the Leiden resolution, and * denotes optional content that indicates the cell boundary correction status. If the * part is absent, the .h5ad file contains clustering results based on the feature expression matrix of cell nuclei-covered regions. However, if the * is replaced by .adjusted, the .h5ad file records clustering results based on the feature expression matrix of cell nuclear expression distance, considering a 10-pixel radius until another cell boundary is encountered. In a typical SAW pipeline, the cell clustering analysis is processed with the Leiden resolution of 1.0. Only available when the image was processed in SAW pipelines since the cell location information is derived from microscope images.

Image Processing Input Files

SAW embeds automated image processing algorithms to stitch image tiles, identify the boundaries of tissue and cells, and detect tracklines on the Stereo-seq chip for aligning the image with the feature expression matrix that also has tracklines. In cases where the tracklines cannot be detected or the tissue/cell boundaries are vague, you may need to manually outline or align.

The Image Processing module is designed to address the need for outlining tissue or cells manually or importing results from other tools, as well as manually aligning an image to the feature expression matrix. This manual intervention can be performed on the original microscope image or the SAW automated image detection algorithms processed data.

File extension	Description
.tar.gz	The compressed TAR file saves the original microscope images and the QC information. Generated from Tools -> Stereo-seq Image QC.
.tif or .tiff	The TIFF format image. Microscope stitched image.
.stereo	Visualization manifest file in JSON format that includes information about where to find SAW automated image processing results and feature expression matrix. Unzip the SAW-generated bundle file (visualization.tar.gz) to find this .stereo manifest file. If the .stereo file is the Image Processing module input, at least a compressed image TAR file (.tar.gz) and a feature expression matrix file (.gef) must be available in the directory. Both paths need to be recorded in the manifest file for easy reference. The image .tar.gz file in the SAW output bundle file saves the auto image segmentation and registration records.

Depending on the content of the input files, the expected result from Image Processing might be different. More details are described in Navigation for Image Processing.