In MassDynamics/lfq_processing: Processes Proteomics Files for Quantification.

knitr::opts_chunk$set(fig.cap = NULL, fig.path = params$output_figure, echo=FALSE, eval=TRUE, dev = "svg")

library(data.table)
library(ggplot2)
library(ggrepel)
library(GGally)
library(umap)
library(FactoMineR)
library(factoextra)
library(corrplot)
library(viridis)
library(ggpubr)
library(Hmisc)
library(plotly)
library(stringr)
library(bit64)

num_files <- expdes[, .N]
run_per_condition <- expdes[, .(countRepMax = .N), by = .(experiment)]
setnames(run_per_condition, "experiment", "condition")

# for fractions, create file name from mqExperiment and Fraction
if (!("file_name" %in% colnames(expdes))){
  expdes$file_name = paste(expdes$experiment, " - ", expdes$Replicate)
}

prot_int_rep <- merge(
  run_per_condition,
  prot_int[Imputed == 0, .(Repcount = .N), by = .(id, condition)],
  by = c("condition")
)
prot_int_rep[, repPC := Repcount/countRepMax]
prot_id_in_a_cond <- prot_int_rep[repPC >= 0.5, unique(id)]
prot_int[, Valid := 0L]
prot_int[id %in% prot_id_in_a_cond, Valid := 1L]
rm(prot_id_in_a_cond, prot_int_rep)

if ("file_name" %in% colnames(prot_int)){
  filenames <- prot_int[!(is.na(file_name)), unique(run_id)]
} else {
  filenames <- prot_int[, unique(run_id)]
}

int_corr_dt <- dcast.data.table(
  prot_int[Imputed == 0],
  id ~ condition + Replicate,
  value.var = "log2NInt",
  sep = "_"
)

# must have >4 observations for spearman correlation to run
test = try(rcorr(as.matrix(int_corr_dt[, 2:ncol(int_corr_dt)])), silent = TRUE)

if (class(test) != "try-error") {

  DT_corMatrix <- rcorr(as.matrix(int_corr_dt[, 2:ncol(int_corr_dt)]))
  DT_corMatrix <- DT_corMatrix$r
  DT_corMatrix[DT_corMatrix <= -1] = -1
  DT_corMatrix[DT_corMatrix >= 1] = 1
  #corrplot(DT_corMatrix, type = "upper", title = "Correlation Matrix", tl.cex = 0.5, mar = c(1,0,1.5,0))
  if (num_files > 10) {
    corrplot(DT_corMatrix, type = "upper", title = "Correlation Matrix - Proteins", tl.cex = 0.5, mar = c(1,0,1.5,0))

  } else {
    corrplot(DT_corMatrix, type = "upper", title = "Correlation Matrix - Proteins", tl.cex = 0.5, mar = c(1,0,1.5,0), addCoef.col = "white", number.cex = .5)

  }

}