Preserves the MycoHalo argument names. For MI with set_id metadata this
delegates to assess_experiment() so each treatment uses its own control.
For other outcomes, or a legacy single-set table, it fits the original
plate-aware comparison with lmerTest Satterthwaite inference instead of
the old ad hoc degrees of freedom. Multi-set assessment of MI should use
assess_experiment() directly. Never compare shared-denominator MC rows as
independent observations.
Usage
compare_melanization(
colonies,
response = "MI_mean",
group = "treatment",
plate = "plate_id",
reference = NULL,
method = c("auto", "mixed", "plate_means"),
by = NULL,
control = NULL,
require_qc = FALSE
)Arguments
- colonies
Colony table.
- response
Numeric response column, default MI_mean.
- group
Treatment column.
- plate
Plate ID column.
- reference
Reference treatment; defaults to control if available.
- method
auto, mixed, or plate_means. auto chooses mixed with at least three plates per group, otherwise an explicit plate-mean model.
- by
Extra control-matching columns for matched MI assessment.
- control
Optional control label, alias of reference for matched MI.
- require_qc
Require explicit image QC approval for matched MI analysis. Default FALSE warns on unconfirmed recorded QC.