estimateGridSearchDecoyFDR: Estimate feature grid search FDR

In CCprofiler/CCprofiler: Complex feature detection using SEC protein chromatograms.

Description

Estimate feature grid search FDR by decoy counting.

Usage

estimateGridSearchDecoyFDR(complex_features_list,
  grid_search_params = c("corr", "window", "rt_height",
  "smoothing_length", "min_feature_completeness",
  "min_hypothesis_completeness", "min_subunits", "min_peak_corr",
  "min_monomer_distance_factor"))

Arguments

complex_features_list	data.table containing filtered complex feature results.
grid_search_params	Character vector of column names to report with the statistics for the dataset. Should contain all parameters that are of interest in the grid search.

Value

A data.table with FDR estimations for every parameter combination specified.

Examples

## Generate example data
complexFeaturesGrid <- list(exampleComplexFeatures)

complexFeaturesGridFiltered <- filterGridSearchResults(complexFeaturesGrid,
                                                       peak_corr_cutoffs = c(0.5,0.75,0.9),
                                                       feature_completeness_cutoffs = c(0,0.5,1),
                                                       hypothesis_completeness_cutoffs = c(0.5,1),
                                                       n_subunits_cutoffs =c(2,3,4),
                                                       monomer_distance_cutoffs = c(1,2),
                                                       remove_decoys=FALSE)

## Calculate the FDR statistic for every grid search result
# Here we only performed grid search of different filter cutoffs, so only those are
# specified as grid_search_params

gridStats <- estimateGridSearchDecoyFDR(complexFeaturesGridFiltered,
                                        grid_search_params =c("min_feature_completeness",
                                                              "min_hypothesis_completeness",
                                                              "min_subunits",
                                                              "min_peak_corr",
                                                              "min_monomer_distance_factor"))
## Inspect the oputput
gridStats

CCprofiler/CCprofiler documentation built on May 19, 2021, 7:40 p.m.