Convert native differential-expression results to a stable table

The helper does not replace native result objects. It extracts the small set of fields needed by downstream enrichment, drug-query, and meta-analysis functions while retaining feature identifiers explicitly. For edgeR likelihood-ratio and quasi-likelihood F tests, statistic is a directional signed square-root score for ranking only. Such tests do not expose a coefficient standard error, so standard_error remains NA and meta_collect() rejects them instead of manufacturing an invalid value. A DESeq2 likelihood-ratio statistic is retained in the table but marked as non-directional; de_ranks() requires column = "effect" for that result, or a Wald result when a directional test statistic is desired. Generic data-frame F and LR columns are likewise retained as non-directional because their degrees of freedom cannot be inferred safely.

Usage

de_table(
  result,
  coef = NULL,
  feature_column = NULL,
  effect_column = NULL,
  standard_error_column = NULL,
  statistic_column = NULL,
  p_value_column = NULL,
  adjusted_p_value_column = NULL
)

Arguments

result: A DESeqResults, edgeR DGELRT/DGEExact, limma/dream MArrayLM, or data-frame-like result.
coef: Coefficient name or index for an MArrayLM with more than one coefficient.
feature_column, effect_column, standard_error_column, statistic_column: Optional feature, effect, standard-error, and statistic column names for a generic data frame. Common backend names are detected automatically.
p_value_column, adjusted_p_value_column: Optional raw and adjusted p-value column names for a generic data frame. Common backend names are detected automatically.

Value

A data frame with feature_id, effect, standard_error, statistic, p_value, and adjusted_p_value columns.