Skip to contents

Preprocess CACTI-S Segments

Source: R/step_s_preprocess.R
cacti_s_preprocess.Rd

Performs QC, filtering, and normalization

Usage

cacti_s_preprocess(
  file_raw_counts,
  file_out,
  file_out_meta,
  filter_mode = c("cpm_threshold", "match_overlap_count", "prop"),
  filter_bed = NULL,
  prop_top = 0.2,
  min_cpm = 1,
  min_prop = 0.2
)

Arguments

file_raw_counts

Path to featureCounts output (from cacti_s_count).

file_out

Output path for the processed phenotype file.

file_out_meta

Output path for the meta file of processed phenotype.

filter_mode

Strategy to select segments: "cpm_threshold" (default), "match_overlap_count", or "prop".

filter_bed

(Required for "match_overlap_count") Path to a BED file (e.g. peaks or QTL regions).

prop_top

(For "prop") Proportion of segments to keep (0.0 to 1.0). Default 0.2.

min_cpm

(For "cpm_threshold") Minimum CPM threshold (default 1).

min_prop

(For "cpm_threshold") Minimum proportion of samples (default 0.2).

Value

Invisibly returns the processed data frame. Writes file_out to disk.

Details

Processing Steps:

  1. QC: Remove segments with 0 reads in all samples.

  2. CPM: Transform counts to log2(CPM + 0.5).

  3. Filter: Subset segments based on filter_mode.

  4. Standardize: Z-score normalization per feature (Mean=0, SD=1).

  5. RankNorm: Inverse Normal Transform per sample.

Filtering Logic:

  • cpm_threshold (Default): Keeps segments with peak intensity > min_cpm in > min_prop of samples. Standard method to remove noise.

  • match_overlap_count: Determines $N$, the number of segments overlapping regions in filter_bed. Then keeps the top $N$ segments by average intensity.

  • prop: Keeps the top prop_top proportion of segments by average intensity.

Examples

if (FALSE) { # \dontrun{
# Locate the bundled test data
filecounts <- system.file("extdata/test_results", "test_raw_counts.txt", package = "cacti")
filepeaks  <- system.file("extdata", "test_peaks.bed", package = "cacti")

# Define output file
fileout <- "inst/extdata/test_results/test_pheno_norm.txt"
fileout <- "inst/extdata/test_results/test_pheno_norm_meta.txt"

# Check if files exist (only true if package is installed)
if (nchar(filecounts) > 0) {

  # --- Example 1: Default Filtering (CPM Threshold) ---
  cacti_s_preprocess(
    file_raw_counts = filecounts,
    file_out = fileout,
    file_out_meta = fileout_meta,
    filter_mode = "cpm_threshold",
    min_cpm = 10,
    min_prop = 0.5
  )
  head(read.table(fileout, header = TRUE))

  # --- Example 2: Match Overlap Count ---
  # Logic: If 2 segments overlap peaks, keep the top 2 segments by intensity.
  cacti_s_preprocess(
    file_raw_counts = filecounts,
    file_out = fileout,
    file_out_meta = fileout_meta,
    filter_mode = "match_overlap_count",
    filter_bed = filepeaks
  )
  head(read.table(fileout, header = TRUE))
}
} # }