Processing QC-failed Image

Several microscope images fail quality control when trackline detection is unsuccessful, but they still need to be used for SAW realign. This tutorial outlines procedures for processing microscope images that failed QC checks, specifically focusing on registration and segmentation steps.

Prerequisites

1. Determining the QC-failed image that needs to be processed

   • StereoMap Tools-Image QC shows the QC result

   • Image(s) have been input by --image-tar or --image in SAW count, but the pipeline output HTML summary report does not contain image information.

2. The feature expression matrix has been generated (.stereo file obtained)

Ensure that the SAW count has been completed to create the feature expression matrix. This matrix is critical for aligning spatial data with the microscope image during morphology registration.

• If SAW count inputs do not contain the QC-failed image, the expression matrix lacks alignment with the microscope image. You need to perform morphology registration in Image processing Step 2.

• If the QC-failed image was input into SAW count, the pipeline will still place QC-failed images into /outputs/visualiztion.tar , but will not perform automatic registration. You also need to perform morphology registration in Image processing Step 2.

Morphology registration

When QC failed, SAW automated registration or StereoMap Feature Point registration becomes unavailable. Hence, Morphology registration becomes essential to align the expression matrix with the microscope image. Two functions are helpful:

• Adjust the bin size

  • A smaller bin size enhances detail but may increase computational load, while a larger bin size simplifies the image for coarse alignment.

• Trackline overlap

  • Even if trackline detection fails might be due to insufficient exposure, residual trackline patterns can still be compared with a reference template for fine-tuning registration.

  • Click"Chip tracklines" to display tracklines derived from the spatial feature expression matrix. Overlay the QC-failed image’s tracklines with the matrix's one.

  • Use a small bin size to visually inspect and refine the alignment between the microscope image and the reference trackline template. Adjust rotation and scaling parameters to maximize overlap between the two trackline patterns.

Choose tissue & cell segmentation methods

Depending on your analysis requirements, you can decide tissue & cell segmentation methods after morphology registration. These diagrams explain SAW realign whether to use manual/third-party segmentation results or enable pipeline auto-segmentation.

Timing for enabling SAW pipeline auto-segmentation

If you do not complete Step 3 Tissue segmentation & Step 4 Cell segmentation in StereoMap Image processing, SAW realign will activate built-in automatic tissue and cell segmentation in the pipeline.

If you complete Step 3 Tissue segmentation but skip Step 4 Cell segmentation in Image processing, SAW realign will activate built-in cell segmentation in the pipeline.

If you complete both manual/third-party tissue & cell segmentation in Image processing, SAW realign directly uses segmentation results in the pipeline.

How to upload third-party segmentation mask(s)

• Export the registered TIFF from StereoMap and use third-party tools to define tissue and cell masks.

• Import third-party tool segmentation masks back into StereoMap and record them in the output processed TAR.GZ.

SAW realign

cd /saw/runs

saw realign \
    --id=Adjuated_Demo_Mouse_Brain \
    --sn=SS200000135TL_D1 \
    --count-data=/path/to/previous/SAW/count/task/folder \
    --realigned-image-tar=/path/to/realigned/image/tar