Skip to content

quick-start.md

Latest commit

 

History

History
51 lines (37 loc) · 2.94 KB

quick-start.md

File metadata and controls

51 lines (37 loc) · 2.94 KB
Raw

Quick Start

PRISM is the post-stitching half of the pipeline: readout → gene_calling → cell_segmentation → analysis. For the pipeline overview and diagram, see the main README; for sample-data downloads, see Data Sources.

Install the package first — see the Installation Guide. All commands below run from the repository root.

Option A: Start from Stitched Images (Recommended for Beginners)

Prerequisites:

Workflow Steps:

  1. Spot Detection / Readout — edit base_dir (and, if needed, channel_files) in configs/readout.yaml, then run:

    python scripts/readout.py <RUN_ID>

    Outputs readout/position.csv and readout/intensity.csv.

  2. Gene Calling — set BASE_DIR and RUN_ID at the top of scripts/gene_calling.py (or pass a config file), then run:

    python scripts/gene_calling.py

    Outputs readout/mapping.csv (gene assignments + confidence) and readout/intensity_corrected.csv. The default method is GMM / codebook-GMM. For an interactive walkthrough, use notebooks/readout_gene_calling_2D.ipynb or notebooks/gene_calling_GMM.ipynb.

  3. Cell Segmentation — segment nuclei from the DAPI channel:

    python scripts/segment_dapi.py <RUN_ID>

    Outputs nucleus centroids (segmented/dapi_centroids.csv) and a label image.

  4. Cell-by-gene matrix & cell typing — assign RNA spots (mapping.csv) to their nearest nucleus and build the expression matrix in notebooks/cell_typing_and_analysis.ipynb.

Expected Outputs:

  • readout/position.csv, readout/intensity.csv: detected RNA spots (coordinates + per-channel intensities)
  • readout/mapping.csv: spots with gene assignments and confidence
  • segmented/dapi_centroids.csv: nucleus centroids
  • cell-by-gene expression matrix: built in the cell-typing notebook

Option B: Start from Raw Unstitched Images (Full Pipeline)

PRISM does not ship an in-tree image-processing layer. For raw → stitched output (focal stacking, illumination correction, registration, pcorr_bigstitcher / MIST stitching, AIRLOCALIZE), use the companion spatial_img_core package (not yet public — request access at huanglab111@gmail.com). Once stitched images land under <RUN_ID>_processed/stitched/, continue with Option A above.

Next Steps