Stereo Sequencing Data Processing

This document provides an overview of the script designed to process spinal cord stereo sequencing data. The main steps undertaken by the script are:

1. Visualization and Coordination:

• Objective: Visualize captured RNA sequences to identify the spatial region associated with each sample.
• Procedure:
  • For each sample, determine and record the boundaries in terms of x_min, x_max, y_min, and y_max. These boundaries set the stage for subsequent data cropping operations.

2. Data Cropping:

• Objective: Crop sections of multiple samples sequenced on the same chip based on the predetermined coordinates.
• Procedure:
  • The script processes two types of data:
    1. Bin1 data (with the .gef extension).
    2. Cell bin data (also with the .gef extension).
  • After processing, the cropped results are saved as .h5ad files, which are suitable for downstream analysis.

3. Create a tissue segmentation using labelme based on bin1 plot, here in the bin1 plot exported image, each pixel corresponding to 1 unit in the chip. The labelme segmentation coordiante could be directly applied to the gene matrix cutting.

4. Voronoi Generation based on cell bin center

• Objective: Redefine spatial regions based on cell bin centers and re-allocate the Bin1 data into these newly defined cell bins.
• Procedure:
  • The script creates new Voronoi tessellations using the centers of the cell bins.
  • Regions outside the tissue segmentation will be filtered out.

5. Obtaining offset coordinate for all sections.

• Offset the cell bin json file coordinate based on the bin1 h5ad minx and min y that all the json polygon could be coupled with the bin1 cropped image.
• Offset the h5ad bin1 file for following up processing.
• then we need to get the minx and miny from the cropped bin1 h5ad files

5. Cell Cut

• The Bin1 data is then rebinned according to these new cell bin divisions.
• This process, known as "cell cut," ensures a more accurate representation of the spatial RNA data.

---

Note: Users are advised to familiarize themselves with the terms and data formats (such as .gef and .h5ad) to ensure smooth processing and accurate interpretation of the results.