In MassDynamics/lfq_processing: Processes Proteomics Files for Quantification.

knitr::opts_chunk$set(fig.cap = NULL, fig.path = params$output_figure)

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)
}


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

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


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)

pca_dt <- pept_int[Valid == 1, .(
  id,
  condition,
  Replicate,
  log2NInt,
  run_id
)]
pca_dt <- dcast(pca_dt, condition + Replicate ~ id, value.var = "log2NInt")


num_dimensions = min(pca_dt[, .N-1], 10)
res.pca <- PCA(pca_dt[, 3:ncol(pca_dt)], graph = FALSE, ncp = num_dimensions)
eig.val <- get_eigenvalue(res.pca)
eig.val <- data.table(dims = rownames(eig.val), eig.val)

scree_plot <- ggplot(eig.val[1:num_dimensions], aes(x=reorder(dims, -`variance.percent`), y=`variance.percent`)) +
  geom_bar(stat="identity", fill = "skyblue2") +
  theme_minimal() +
  geom_text_repel(aes(label=(round(`variance.percent`,1))), direction = 'y') +
  scale_x_discrete("PCA components") +
  scale_y_continuous("% Variance") +
  ggtitle("Peptides - PCA Screeplot")

ggplotly(scree_plot, tooltip = c("y")) %>% config(displayModeBar = T, 
                                                  modeBarButtons = list(list('toImage')),
                                                  displaylogo = F)