Benchmark annotation engines on the same input — benchmark_annotation

Runs one or more annotation engines on the same expression and metadata inputs, then summarizes coverage, ambiguity, and provenance for comparison.

Usage

benchmark_annotation_engines(
  expr,
  meta = NULL,
  species = c("auto", "human", "mouse"),
  annotation_preset = NULL,
  engines = c("none", "biomart", "orgdb", "ensdb", "hybrid"),
  fields = NULL,
  strip_version = TRUE,
  biomart_version = 102,
  biomart_host = NULL,
  biomart_mirror = NULL,
  assay_name = NULL,
  gene_id_col = NULL,
  sample_col = "sample",
  output_file = NULL,
  coverage_file = NULL,
  verbose = TRUE
)

Arguments

expr: Expression table with gene_id first, or a SummarizedExperiment-like object.
meta: Metadata table with sample first. Leave NULL when expr is a SummarizedExperiment.
species: Either "auto", "human", or "mouse".
annotation_preset: Optional preset that fixes a reproducible annotation configuration. Supported values are "human_v102", "mouse_v102", "human_tpm_v102", "mouse_tpm_v102", "human_count_v102", and "mouse_count_v102".
engines: Character vector of annotation engines to compare.
fields: Optional annotation fields to benchmark.
strip_version: Whether to remove Ensembl version suffixes.
biomart_version: Fixed Ensembl release used by the biomaRt backend.
biomart_host: Optional explicit Ensembl host.
biomart_mirror: Optional Ensembl mirror name.
assay_name: Optional assay name when expr is a SummarizedExperiment.
gene_id_col: Optional row-data column containing Ensembl IDs when expr is a SummarizedExperiment.
sample_col: Metadata column to use as sample when expr is a SummarizedExperiment.
output_file: Optional CSV path for the benchmark summary.
coverage_file: Optional CSV path for long-format coverage results.
verbose: Whether to emit progress messages.

Value

An expranno_benchmark object.