Denoised Per-Detector Variability (MAD)

As assess.variability(), but returns the full per-detector robust SD (MAD) profile across a fluorophore's variant spectra, rather than a single summary value. Detectors where the reference spectrum falls below peak.height are zeroed out before computing MAD, so that off-peak noise does not contribute to the profile.

Usage

assess.variability.mad(variant.list, peak.height = 0.1)

Arguments

variant.list

Named list of spectral variant matrices (one per fluorophore), as returned in the variants element of get.spectral.variants(). Row 1 of each matrix is treated as the reference spectrum; rows 2+ are treated as variants.

peak.height

Numeric in [0, 1], default 0.1. Detectors where the reference spectrum falls below this fraction of that fluorophore's own peak value (peak.height * max(reference.spectrum)) are set to zero before MAD is computed. Relative to the peak rather than an absolute cutoff, so behavior is unchanged whether variant.list is L-infinity or L2 normalized.

Value

A numeric matrix with one row per fluorophore and one column per detector, giving the denoised per-detector MAD of the variant spectra.