tests/testthat/test_filtering.R

In peterblattmann/SWATH2stats: Transform and Filter SWATH Data for Statistical Packages

context("data processing")

test_that("filtering", {
  data(OpenSWATH_data, package="SWATH2stats")
  data <- OpenSWATH_data
  data(Study_design, package="SWATH2stats")

  data <- reduce_OpenSWATH_output(data)
  
  # test if checking for missing columns works
  data2 <- data
  data2$ProteinName <- NULL
  expect_warning(reduce_OpenSWATH_output(data2),"These columns are missing from the data:ProteinName")
  
  # sample_annotation
  data <- sample_annotation(data, Study_design)
  
  ## assess_decoy_rate
  assess_decoy_rate(data)
  expect_that(assess_decoy_rate(data), shows_message("Number of non-decoy identifiers: 273"))
  expect_that(assess_decoy_rate(data), shows_message("Number of decoy identifiers: 11"))
  expect_that(assess_decoy_rate(data), shows_message("Decoy rate: 0.0403"))

  # test if no decoys present
  data2 <- data
  data2$decoy <- NULL
  expect_error(assess_decoy_rate(data2), "There is no decoy column in the table")
  # test if no decoys present
  data2 <- data
  data2$FullPeptideName <- NULL
  expect_error(assess_decoy_rate(data2))
  
  ## filter_mscore
  data.filtered.mscore <- filter_mscore(data, 0.01)
  data.filtered.mscore <- filter_mscore(data, 0.001)
  data.filtered.mscore <- filter_mscore(data, 0.0001)

  expect_that(filter_mscore(data, 0.01), shows_message("Dimension difference: 931, 0"))
  expect_that(filter_mscore(data, 0.001), shows_message("Dimension difference: 1068, 0"))
  expect_that(filter_mscore(data, 0.0001), shows_message("Dimension difference: 1166, 0"))

  # filter mscore_requant
  data.filtered.mscore <- filter_mscore_freqobs(data, 0.01, 0.8)

  expect_that(filter_mscore_freqobs(data, 0.01, 0.8), shows_message("Treshold, peptides need to have been quantified in more conditions than: 4.8"))
  expect_that(filter_mscore_freqobs(data, 0.01, 0.8), shows_message("Fraction of peptides selected: 0.42"))
  expect_that(filter_mscore_freqobs(data, 0.01, 0.8), shows_message("Dimension difference: 1323, 0"))

  expect_message(filter_mscore_freqobs(data, 0.01), "Treshold, peptides need to have been quantified in more conditions than: 0")
  
  # filter mscore_conditions
  data.filtered.mscore <- filter_mscore_condition(data, 0.01, 3)
  expect_that(filter_mscore_condition(data, 0.01, 3), shows_message("Fraction of peptides selected: 0.47"))
  expect_that(filter_mscore_condition(data, 0.01, 3), shows_message("Dimension difference: 1209, 0"))

  # filter on_fdr
  expect_that(filter_mscore_fdr(data, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05),
  shows_message("been removed from the returned data"))
  
  # test if remove decoy works
  data2 <- filter_mscore_fdr(data, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05, rm_decoy=FALSE)
  data3 <- filter_mscore_fdr(data, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05, rm_decoy=TRUE)
  
  expect_message(filter_mscore_fdr(data, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05, rm_decoy=FALSE),
                 "The decoys have NOT been removed from the returned data")
  expect_true(length(grep("DECOY", data2$ProteinName)) > 1)
  expect_false(length(grep("DECOY", data3$ProteinName)) > 1)
  
  
  # test that error is returned if lowest m_score is from decoy
  data2 <- data
  data2 <- data2[data2$m_score != 0,]
  data2[data2$decoy == 1,"m_score"] <- data2[data2$decoy == 1,"m_score"]*1e-20
  expect_true(is.na(mscore4protfdr(data2, FFT = 0.7, fdr_target = 0.001)))
  expect_error(filter_mscore_fdr(data2, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05), 
               "The overall_protein_fdr_target cannot be reached in this dataset")
  
  # test that local FDR is only calculated if enought decoys for each run
  data2 <- data
  data2 <- data2[data2$m_score != 0,]
  data2[data2$decoy == 1 & data2$m_score == 2,"m_score"] <- 0.01

  expect_message(filter_mscore_fdr(data, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05), 
               "Individual run FDR quality of the peptides was not calculated\nas not every run contains a decoy")
  expect_message(filter_mscore_fdr(data2, FFT = 0.7, overall_protein_fdr_target=0.1, upper_overall_peptide_fdr_limit=0.05), 
                 "Individual run FDR quality of the peptides selected from")
  
              
})

test_that("count_analytes", {
  data(OpenSWATH_data, package="SWATH2stats")
  data <- OpenSWATH_data
  
  analysis <- count_analytes(data)
  expect_that(analysis[analysis$run_id == "0_210","FullPeptideName"], equals(261))
  expect_that(analysis[analysis$run_id == "0_221","FullPeptideName"], equals(271))
  
  analysis2 <- count_analytes(data[data$m_score < 1e-4,])
  expect_that(analysis[analysis$run_id == "0_177","transition_group_id"], equals(388))
  expect_that(analysis[analysis$run_id == "0_29","ProteinName"], equals(10))
})