Skip to contents

Construct a SummarizedExperiment object from raw expression table

Source: R/QC.R
rawdata2se.Rd

This function reads a raw expression table, performs feature-level filtering, optional outlier masking, missing-value imputation, and stability filtering, and returns a SummarizedExperiment object together with diagnostic gene lists.

Usage

rawdata2se(
  exp_file,
  obs_col = "Genes",
  raw_prefix = "auto",
  enable_detect_outlier_gene = FALSE,
  fc_threshold = 5,
  min_valid_groups = 0,
  frac_NA_threshold = 0.5,
  min_stable_groups = 0,
  cv_threshold = 0.5
)

Arguments

exp_file

Path to the raw expression table file. The file must contain one column specifying feature identifiers and multiple columns containing raw expression values.

obs_col

Name of the column in exp_file used to define feature identifiers. Only the substring before the first semicolon (;) will be used as the feature name.

raw_prefix

Character string used to identify raw expression columns. If \'auto\' (default), raw intensity columns will be detected automatically from common raw file suffixes such as \'raw\', \'mzML\', \'mzXML\', and similar filename-style endings.

enable_detect_outlier_gene

Logical flag indicating whether to mask extreme outlier measurements within each condition based on fold-change from the median.

fc_threshold

Numeric threshold for fold-change–based outlier detection. Values exceeding this fold change relative to the condition median will be set to NA when outlier detection is enabled.

min_valid_groups

Minimum number of conditions in which a feature must pass the missing-value filter to be retained. Features with fewer valid conditions are removed.

frac_NA_threshold

Maximum allowed fraction of missing values within a condition for that condition to be considered valid for a given feature.

min_stable_groups

Minimum number of conditions in which a feature must show low variability (coefficient of variation below cv_threshold) to be retained.

cv_threshold

Numeric threshold on coefficient of variation (CV) used to define feature stability within each condition.

Value

A list with class "RawDataSE" containing:

  • se: A SummarizedExperiment object storing raw intensities, imputed intensities, CPM-normalized values, and z-scored expression.

  • un_stable_gene: A data frame containing the original input rows of features removed due to high variability across conditions.

  • missing_gene_df: A data frame containing the original input rows of features removed due to excessive missing values.