plot_rank_metrics — plot_rank_metrics • clustOpt

Create a publication-quality plot showing individual metric ranks and mean rank across clustering resolutions.

Usage

plot_rank_metrics(
  rank_results,
  method = "rank",
  highlight_best = TRUE,
  base_size = 12
)

Arguments

rank_results: A data.frame output from suggest_resolution.
method: Character; "rank" (default) for direct rank aggregation, or "curvature" for curvature-based ranking. The curvature method requires at least 3 resolutions.
highlight_best: Logical; if TRUE (default), highlight the resolution with the lowest mean rank with a vertical dashed line.
base_size: Numeric; base font size for theme. Default is 12.

Value

A ggplot object

Details

The plot shows:

Individual metric ranks as colored lines/points
Mean rank as a bold black line
Lower ranks indicate better performance (y-axis is reversed)
The best resolution (lowest mean rank) is optionally highlighted

Uses a colorblind-friendly palette (Okabe-Ito) for metric colors.

See also

suggest_resolution, plot_mean_rank

Examples

if (FALSE) { # \dontrun{
cv_results <- clust_opt(seurat_obj, subject_ids = "donor_id")
rankings <- suggest_resolution(cv_results)
plot_rank_metrics(rankings)
plot_rank_metrics(rankings, method = "curvature")
} # }