Reads the editable project metadata, applies the fungal or control layout per row, writes QC images and saves measurements. Review flags are reset for this new analysis; coefficients are withheld until finish_experiment().
Usage
run_experiment(
project,
n_spots = NULL,
spot_radius_mm = 0.5,
...,
default_centre = "fungus"
)Arguments
- project
A funmel_project, its current folder, project.rds, or a saved analysis beside project.rds. Use load_experiment() after moving a project.
- n_spots
Optional number of circular spots per target colony; NULL uses whole-colony pixels only. Spot and whole-colony results are kept separate.
- spot_radius_mm
Radius of each optional circular ROI (mm).
- ...
Other analyze_plate() settings, such as dish_diameter_mm and max_dim.
- default_centre
Segmentation fallback for unlabeled images, default fungus. Use none for control-only images. Per-image centre_model and known control labels take precedence. No treatment or control identity is guessed.
Details
Dish diameter and any supplied pixel scale must be finite positive numeric scalars. Invalid settings stop before image processing or changes to project QC flags. Missing project metadata must be restored or its recorded path corrected. A batch with failed images saves error reports and warns; it does not report success. Inspect result$failed and resolve every failed image before finalization. Image completion is separate from colony completeness. run_summary.csv reports finite target MI counts; measurement_issues.csv identifies missing colonies and incomplete spot sampling. metadata_issues.csv lists design fields and affected plates, while layout_assignments.csv records the actual layout and its source.