Classify genes by multi-round monotonic time-course patterns

This function classifies genes based on monotonic temporal trends using an iterative Spearman correlation strategy across progressively trimmed time windows. At each round, genes exhibiting significant positive or negative monotonicity are assigned to Up or Down groups corresponding to the round at which the trend first becomes detectable.

Usage

classify_timecourse_by_round(
  mat,
  rho_cut = 0.75,
  min_len = 4,
  trim_side = c("best", "head", "tail")
)

Arguments

mat

A numeric matrix with genes in rows and time-ordered samples in columns. Row names must correspond to gene identifiers.

rho_cut

Numeric scalar specifying the absolute Spearman correlation threshold for calling a monotonic trend. Default is 0.75.

min_len

Integer specifying the minimum number of samples required to perform trend classification. Default is 4.

trim_side

Character string specifying how to trim the time window at each iteration. One of:

Value

A data.frame with one row per gene and the following columns:gene,gene_group,type

Details

The time window is shortened by removing one boundary sample (head, tail, or automatically selected) after each round, until fewer than min_len samples remain.