tests/testthat/test_plots.R

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

context("plotting")

test_that("correlation plot", {
  data(OpenSWATH_data, package="SWATH2stats")
  data(Study_design, package="SWATH2stats")
  data.FDR<-sample_annotation(OpenSWATH_data, Study_design)
  
  
  correlations <- plot_correlation_between_samples(data.FDR)
  expect_that(correlations[correlations$Var1 == "Hela_Control_1" & correlations$Var2 == "Hela_Control_1","value"], equals(1))
  expect_that(correlations[correlations$Var1 == "Hela_Treatment_1" & correlations$Var2 == "Hela_Treatment_1","value"], equals(1))
  
  correlations <- plot_correlation_between_samples(data.FDR, label = FALSE)
  expect_that(correlations[correlations$Var1 == "Hela_Control_1" & correlations$Var2 == "Hela_Control_1","value"], equals(1))
  
  
  data1 <- data.FDR[data.FDR$Condition == "Hela_Control" & data.FDR$BioReplicate == 1 & data.FDR$decoy == 0, c("transition_group_id", "Intensity")]
  data2 <- data.FDR[data.FDR$Condition == "Hela_Treatment" & data.FDR$BioReplicate == 1 & data.FDR$decoy == 0, c("transition_group_id", "Intensity")]
  data3 <- merge(data1, data2, by="transition_group_id")
  
  cor.p <- cor(data3$Intensity.x, data3$Intensity.y, method="pearson")
  cor.s <- cor(data3$Intensity.x, data3$Intensity.y, method="spearman")
  
  expect_that(correlations[correlations$Var1 == "Hela_Control_1" & correlations$Var2 == "Hela_Treatment_1" & correlations$method == "pearson","value"], equals(cor.p))
  expect_that(correlations[correlations$Var1 == "Hela_Treatment_1" & correlations$Var2 == "Hela_Control_1" & correlations$method == "spearman","value"], equals(cor.s))
  
 })

test_that("variation plot", {
  data(OpenSWATH_data, package="SWATH2stats")
  data(Study_design, package="SWATH2stats")
  data.FDR<-sample_annotation(OpenSWATH_data, Study_design)
  
  variation <- plot_variation(data.FDR, label=FALSE)
  variation <- plot_variation(data.FDR)
  variation.data <- variation[[1]]
  
  # test extraction
  test1 <- variation.data[variation.data$transition_group_id == "124947_SGWVKPIIIGVLR_2_run0" & variation.data$Condition == "Hela_Control",]
  test2 <- variation.data[variation.data$transition_group_id == "214321_GTLNLDSYR_2_run0" & variation.data$Condition == "Hela_Treatment",]
  
  val1 <- data.FDR[data.FDR$transition_group_id == "124947_SGWVKPIIIGVLR_2_run0" & data.FDR$Condition == "Hela_Control", "Intensity"]
  val2 <- data.FDR[data.FDR$transition_group_id == "214321_GTLNLDSYR_2_run0" & data.FDR$Condition == "Hela_Treatment", "Intensity"]
  
  expect_true(sum(test1[,c("1","2","3")] %in% val1) == 3)
  expect_true(sum(test2[,c("1","2","3")] %in% val2) == 3)
  
  # test cv calculation
  
  mean.test1 <- mean(as.numeric(test1[,c("1","2","3")]))
  mean.test2 <- mean(as.numeric(test2[,c("1","2","3")]))
  
  expect_true(sd(as.numeric(test1[,c("1","2","3")]))/mean.test1 == as.numeric(test1[,"cv"]))
  expect_true(sd(as.numeric(test2[,c("1","2","3")]))/mean.test2 == as.numeric(test2[,"cv"]))
})


test_that("variation plot vs total", {
  data(OpenSWATH_data, package="SWATH2stats")
  data(Study_design, package="SWATH2stats")
  data.FDR<-sample_annotation(OpenSWATH_data, Study_design)
  
  
  variation <- plot_variation_vs_total(data.FDR)
  variation.data <- variation[[1]]
  
  # test extraction
  test1.total <- variation.data[variation.data$rep == "124947_SGWVKPIIIGVLR_2_run0",]
  test1.rep <- variation.data[variation.data$rep == "124947_SGWVKPIIIGVLR_2_run0 Hela_Control",]
  
  test2.total <- variation.data[variation.data$rep == "214321_GTLNLDSYR_2_run0",]
  test2.rep <- variation.data[variation.data$rep == "214321_GTLNLDSYR_2_run0 Hela_Treatment",]
  
  val1.total <- data.FDR[data.FDR$transition_group_id == "124947_SGWVKPIIIGVLR_2_run0", "Intensity"]
  val2.total <- data.FDR[data.FDR$transition_group_id == "214321_GTLNLDSYR_2_run0", "Intensity"]
  
  val1.rep <- data.FDR[data.FDR$transition_group_id == "124947_SGWVKPIIIGVLR_2_run0" & data.FDR$Condition == "Hela_Control", "Intensity"]
  val2.rep <- data.FDR[data.FDR$transition_group_id == "214321_GTLNLDSYR_2_run0" & data.FDR$Condition == "Hela_Treatment", "Intensity"]

  cv.val1.total <- sd(val1.total)/mean(val1.total)
  cv.val2.total <- sd(val2.total)/mean(val2.total)
  cv.val1.rep <- sd(val1.rep)/mean(val1.rep)
  cv.val2.rep <- sd(val2.rep)/mean(val2.rep)
  
  expect_true(cv.val1.total == as.numeric(test1.total[,"cv"]))
  expect_true(cv.val2.total == as.numeric(test2.total[,"cv"]))
  expect_true(cv.val1.rep == as.numeric(test1.rep[,"cv"]))
  expect_true(cv.val2.rep == as.numeric(test2.rep[,"cv"]))
})