Benchmark Scatter-Match Against Existing AF Assignment Methods

Convenience wrapper that runs assign.af.scatter.match alongside the three existing methods (assign.af.fluorophores, assign.af.residuals, assign.af.joint.cov) on the same test and reference unstained data. For each existing method, the assigned AF variant spectrum is looked up and its cosine similarity to each test cell is computed, allowing direct apples-to-apples comparison with the scatter-match approach.

Usage

benchmark.af.scatter.match(
  test.data,
  ref.data,
  scatter.param,
  spectra,
  af.spectra,
  k = 5L,
  verbose = TRUE
)

Arguments

test.data

Numeric matrix or FCS file path. Test unstained data (cells x channels).

ref.data

Numeric matrix or FCS file path. Reference unstained data (cells x channels).

scatter.param

Character vector of scatter channel names.

spectra

Numeric matrix. Fluorophore spectra (fluorophores x detectors), as used by the existing assign.af.* functions.

af.spectra

Numeric matrix. AF variant spectra (variants x detectors).

k

Integer. Neighbours for scatter-matching. Default 5.

verbose

Logical. Default TRUE.

Value

A list with:

scatter.match

Full output of assign.af.scatter.match.

comparison

Data frame with one row per method and columns: method, mean.cosine, median.cosine, sd.cosine, pct.above.0.9, pct.above.0.95.

per.cell

Data frame with one row per test cell containing cosine similarities from all four methods, for cell-level analysis.