Configuration (YAML Settings)

WormLib uses YAML configuration files to define analysis pipelines. A single config file specifies input/output paths, microscope parameters, channel definitions, and which pipeline steps to run.

—

Configuration Structure

Minimal Example:

input:
  path: ./data/my_images/
  output_directory: ./output/

microscope:
  voxel_size_nm: [1448, 450, 450]  # Z, Y, X in nanometers

channels:
  nuclei:
    name: DAPI
    index: 3

pipeline:
  cell_segmentation: true
  spot_detection: true

Full Example:

input:
  path: ../../data/my_data_folder
  output_directory: ../../output_temp/results

microscope:
  voxel_size_nm: [1448, 450, 450]  # Z, Y, X

channels:
  nuclei:
    name: DAPI
    index: 3
  brightfield:
    name: brightfield
    source: ref_file          # Load from _R3D_REF file
    index: null
  rna:
    - name: mRNA1
      fluorophore: Cy5
      index: 0
      spot_radius_nm: [1409, 340, 340]
      detection_color: red
    - name: mRNA2
      fluorophore: mCherry
      index: 1
      spot_radius_nm: [1283, 310, 310]
      detection_color: blue

segmentation:
  embryo_diameter: 500
  nuclei_diameter: 70

pipeline:
  cell_segmentation: true
  cell_classification: true
  embryo_segmentation: false
  spot_detection: true
  heatmaps: true
  rna_density: true
  line_scan: true

—

Input Section

path — Directory containing microscopy files (.dv, .nd2, .tiff)

output_directory — Where analysis results will be saved

Microscope Section

voxel_size_nm — [Z, Y, X] voxel dimensions in nanometers

For C. elegans 4-cell embryos on DeltaVision: [1448, 450, 450]

Channels Section

Define how to extract channels from your images.

nuclei — Nuclear stain (required for cell segmentation)

  • name — Display name (e.g., “DAPI”)

  • index — Channel index in the image file (0, 1, 2, 3, …)

brightfield — Optional brightfield/transmission image

  • name — Display name

  • source — "ref_file" to load from DeltaVision reference file (_R3D_REF)

  • index — Set to null for brightfield

rna — One or more mRNA channels to detect

For each RNA channel:

  • name — Gene name (e.g., “set3_mRNA”)

  • fluorophore — Dye label (e.g., “Cy5”, “mCherry”)

  • index — Channel index

  • spot_radius_nm — [Z, Y, X] PSF (Point Spread Function) in nm

  • detection_color — Visualization color for detection plots (“red”, “blue”, etc.)

Segmentation Section

embryo_diameter — Expected embryo diameter in pixels (for whole-embryo segmentation)

nuclei_diameter — Expected nucleus diameter in pixels (for nuclear segmentation)

These values help Cellpose optimize segmentation. Typical values:

  • Embryo diameter: 375–500 pixels

  • Nuclei diameter: 30–70 pixels

Pipeline Section

Control which analysis steps run (true/false):

  • cell_segmentation — Segment individual cells

  • cell_classification — Predict cell identity (AB, ABa, EMS, etc.)

  • embryo_segmentation — Segment whole embryo

  • spot_detection — Detect individual mRNA spots

  • heatmaps — Generate grid-based abundance heatmaps

  • rna_density — Analyze RNA density along embryo axis

  • line_scan — Generate line scan intensity plots

Disable steps to save computation time if not needed.

Example Configs

WormLib includes example configurations in config/examples/:

  • one_rna_full.yml — Single channel with all steps enabled

  • two_rna_full.yml — Two RNA channels with cell classification

  • rna_only_spot_detection.yml — Only spot detection, skip segmentation

  • dapi_one_rna_no_brightfield.yml — Nuclear and mRNA, no brightfield

Using Configuration Files

Via Jupyter notebook:

import yaml
with open('config/examples/two_rna_full.yml') as f:
    config = yaml.safe_load(f)

# Access settings
voxel_size = config['microscope']['voxel_size_nm']
output_dir = config['input']['output_directory']

Via command line (future):

python src/wormlib.py --config config/examples/two_rna_full.yml

—

Next Steps