system_genetics/rnaseq/step6_overall_QC/__ - Preloader.R

library(tidyverse)
library(ggfortify)
library(ggprism)
library(limma)
library(biomaRt)
library(PCAtools)
library(gtools)
library(edgeR)
library(ggprism)
library(foreign)


# Global variables
results.dir <- file.path("results.nosync", "RNA-Seq QC")
data.dir <- "Data"
patient.dir <- file.path(data.dir, "Patients")
sample.dir <- file.path(data.dir, "Samples")
expression.dir <- file.path(data.dir, "mRNA - RNA-Seq")


dir.create(results.dir, recursive = TRUE)


####
# Helper functions
####
select.columns.in.order <- function(dataframe, columns) {
  dataframe[, columns]
}

select.rows.in.order <- function(dataframe, rows) {
  dataframe[rows,]
}

getGenedataByEnsemblId38 <- function(ensemblIds, file.location) {
  file.name <- file.path(file.location, "genes_info_hg38.csv")
  if (!file.exists(file.name)) {
    if (!("mart" %in% ls())) {
      assign("mart", useEnsembl(
        biomart = "ENSEMBL_MART_ENSEMBL",
        dataset = "hsapiens_gene_ensembl"
      ))
    }
    gene.list <- getBM(
      filters = "ensembl_gene_id",
      attributes = c(
        "hgnc_symbol",
        "ensembl_gene_id",
        "ensembl_transcript_id",
        "chromosome_name",
        "start_position",
        "end_position",
        "strand",
        "transcription_start_site",
        "transcript_start",
        "transcript_end",
        "external_gene_name"
      ),
      values = as.character(ensemblIds),
      mart = mart
    )
    readr::write_csv(gene.list, path = file.name)
  }
  return(
    readr::read_csv(
    	file.name,
		col_types = readr::cols()
	)
  )
}

#remove.rows.with.count.less.then <- function(dataframe, minRowCount, columns.to.exclude) {
#  dataframe %>%
#  dplyr::filter(
#    rowSums(dplyr::select(., -tidyselect::one_of(columns.to.exclude))) < minRowCount
#  )
#}

limma.voom.convert.column <- function(dataframe, columnname) {
  dataframe %>%
  tibble::column_to_rownames(columnname) %>%
  limma::voom() %>%
  as.data.frame() %>%
  tibble::rownames_to_column(columnname)
}

select.columns.in.order <- function(dataframe, columns) {
  dataframe[, columns]
}

drop.columns.if.all.same.value <- function(dataframe) {
  for (name in colnames(dataframe)) {
    is.all.same <- (dataframe[, name] %>% unique() %>% length()) <= 1
    if (is.all.same) {
      dataframe <- dataframe %>%
        dplyr::select(
          -tidyselect::one_of(name)
        )
    }
  }
  dataframe
}



# Load data
master.Table <- foreign::read.spss(
		file.path(patient.dir, "PRESTO proteogenomics full data sat - ver 7.5.sav")
	) %>%
  as.data.frame() %>%
  dplyr::mutate(
    GenomeScan_ID = stringr::str_trim(GenomeScan_ID),
    gender = forcats::fct_recode(
      Gender,
      female = "f",
      male = "m",
      other = ""
    ),
    age = as.numeric(Age),
    smoking.status = forcats::fct_recode(
      Smoking_status,
      `Ex-smoker` = "ES        ",
      `Current smoker` = "CS        ",
      other = "          "
    )
  )

expression.data <- readr::read_tsv(
		file.path(expression.dir, "20200427_103972-001_rawcounts.txt"),
		col_types = readr::cols()
	)

gene.data <- getGenedataByEnsemblId38(
    ensemblIds = expression.data$Gene,
    file.location = expression.dir
  ) %>%
  dplyr::group_by(hgnc_symbol) %>%
  dplyr::filter(
    dplyr::row_number() == 1,
    !is.na(hgnc_symbol),
    hgnc_symbol != ""
  ) %>%
  dplyr::ungroup() %>%
  dplyr::select(
    hgnc_symbol,
    ensembl_gene_id,
    chromosome_name,
    transcript_start,
    transcript_end
  )


master.Table <- master.Table %>%
  dplyr::mutate(
    Group_simple2 = stringr::str_trim(Group_simple2),
    Group_simple  = stringr::str_trim(Group_simple),
    T_number = as.character(T_number),
    sample.id = stringr::str_trim(PRESTO_ID)
  ) %>%
  dplyr::filter(
    # # According to Niek, I should not include this, for whatever reason
    #!(GenomeScan_ID %in% c(
    #    "T02-01796",
    #    "T02-03095",
    #    "T02-10683",
    #    "T10-18671",
    #    "T12-12036"
    #  )
    # ),
    # 
    # # (According to Niek, don't include) Water Controls
    #!stringr::str_detect(T_number, pattern="Water"),
    # 
    # # (According to Niek, don't include)never smoker controles
    #!(Group_simple2 == "NS_Ctrl"),
    # 
    # # (According to Niek, don't include)ALFA1 patiënten 
    #!(Group_simple == "ALFA"),
    # 
    #stringr::str_trim(Passed_RNAseq_library_prep_QC_Y_N) == "Y",
    GenomeScan_ID %in% colnames(expression.data),
    !is.na(sample.id),
    sample.id != ""
  )

expression.data <- expression.data %>%
  select.columns.in.order(
    c("Gene", master.Table$GenomeScan_ID)
  )