Explore_FireBrowseR

These instructions are for RStudio. In this course we are running RStudio through the Cyverse platform

Package information:

FirebrowsR: https://rdrr.io/github/mariodeng/FirebrowseR/f/vignettes/FirebrowseR.Rmd

Install necessary packages for investigating cancer data stored in the GDAC firehose archive and for data visualization

if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("Biobase")

devtools::install_github("mariodeng/FirebrowseR")

install.packages("ggplot2")

In the console you may be asked if you want to update packages, type ‘a’

go to the library and get the FirebrowseR package

library(FirebrowseR)

Following the FirebrowsR example (from the website above)

##Investigate Breast Cancer and mRNA expression Look at the data in the browser: Navigate to firebrows.org in your browser In the ‘Select cohort’ drop down menu select the option that contains the word ‘breast’ Notice that the cohort associated with ‘breast’ is ‘Breast invasive carcinoma (BRCA)’

Retrieve the data in R: The method Metadata.Cohorts returns all cohort identifiers and their corresponding description. We store these data in cohorts. Within the cohorts we search for (grep) “breast”, yielding to the identifier for breast cancer. note: grep is a ‘shell’ command, not an R-specific command

cohorts = FirebrowseR::Metadata.Cohorts(format = "csv") # Download all available cohorts
cancer.Type = cohorts[grep("breast", cohorts$description, ignore.case = T), 1]
print(cancer.Type)

The return gives us the identifier for breast cancer in this cohort and is the same as the identifier in the browser.

Now, retrieve a list of ‘BRCA’ patients and store the list in brca.Pats ?Samples.Clinical to learn more about the command used here Then we check how many rows and columns of data are stored in brca.Pats

brca.Pats = FirebrowseR::Samples.Clinical(cohort = cancer.Type, format="tsv")
#use base R command `dim` to retrieve the dimensions of `brca.Pats`
dim(brca.Pats)

This output means we retrieved data with 150 rows and 111 columns. The rows correspond to patient identifiers. (as in the Ch 4 table we created)

Scroll to the bottom of the page you opened in your browser that presents the ‘BRCA’ data. What number is displayed in the ‘clinical’ bar? This is how many records we need to pull.

Our R call only returned the first ‘page’ of cases (150). We need to retrieve them all. Use the following

#create `all.Received` and set it to `False` (note the example at Firebrowse R uses the `F` shorthand, we will speall it out to ensure clarity)
all.Received = FALSE
#start at page 1
page.Counter = 1
#the pages have 150 records
page.size = 150
#make an empty list
brca.Pats = list()
#as long as all.Recieved continues to be False keep iterating over the pages and storing the results in the brca.Pats list
while(all.Received == FALSE){
  brca.Pats[[page.Counter]] = Samples.Clinical(format = "csv",
                                               cohort = cancer.Type,
                                               page_size = page.size,
                                               page = page.Counter)
  #use the column names from the first page because subsequent pages don't have col names
  if(page.Counter > 1)
    colnames(brca.Pats[[page.Counter]]) = colnames(brca.Pats[[page.Counter-1]])

  if(nrow(brca.Pats[[page.Counter]]) < page.size){
    all.Received = T
  } else{
    page.Counter = page.Counter + 1
  }
}
brca.Pats = do.call(rbind, brca.Pats)
dim(brca.Pats)

Our output for ‘row’ should now match the number displayed in the ‘Clinical’ bar on the Firebrows R website.

Now subset the data selecting only patients that are dead. This is only to produce a smaller dataset and reduce analysis time. You would only do this is your actual analysis if you only wanted to explore the data from dead patients.

Assign the subset into the same object (brca.Pats = brca.Pats) to overwrite the original file If you want to keep the original file unaltered assign the subset to a different object (e.g. brca.Pats.Sub = brca.Pats)

To view the brca.Pats click on it on the left under ‘Environment’ or use View(brca.Pats) vital_status is a column in brca.Pats We use the $ to tell R we want only that column.

brca.Pats = brca.Pats[ which(brca.Pats$vital_status == "dead"), ]

use ?Samples.mRNASeq to explore this command and understand what is in the parenteses after FirebrowseR::Samples.mRNASeq below

#create a vector of gene's that are known to be diffrentially expressed in breast cancer
diff.Exp.Genes = c("ESR1", "GATA3", "XBP1", "FOXA1", "ERBB2", "GRB7", "EGFR",
                   "FOXC1", "MYC")

#iterate through our `brca.Pats` data set similar to above
all.Found = FALSE
page.Counter = 1
#create an empty list called `mRNA.Exp`
mRNA.Exp = list()
page.Size = 2000 # using a bigger page size is faster
while(all.Found == FALSE){
  mRNA.Exp[[page.Counter]] = FirebrowseR::Samples.mRNASeq(format = "csv",
                                             gene = diff.Exp.Genes,
                                             cohort = "BRCA",
                                             tcga_participant_barcode =
                                               brca.Pats$tcga_participant_barcode,
                                             page_size = page.Size,
                                             page = page.Counter)
  if(nrow(mRNA.Exp[[page.Counter]]) < page.Size)
    all.Found = TRUE
  else
    page.Counter = page.Counter + 1
}
mRNA.Exp = do.call(rbind, mRNA.Exp)
dim(mRNA.Exp)

We want to compare gene expression in samples that have both normal tissue NT and tumor tissue TP samples.

# Patients with normal tissue
normal.Tissue.Pats = which(mRNA.Exp$sample_type == "NT")

# get the patients barcodes
patient.Barcodes = mRNA.Exp$tcga_participant_barcode[normal.Tissue.Pats]

# Subset the mRNA.Exp data frame, keeping only the pre-selected barcodes AND
# having a sample type of NT or TP
mRNA.Exp = mRNA.Exp[which(mRNA.Exp$tcga_participant_barcode %in% patient.Barcodes &
                            mRNA.Exp$sample_type %in% c("NT", "TP")), ]

Go to the library and ‘check out’ the package ggplot2 (installed above)

library(ggplot2)

Plot a box plot of the expression levels for each gene. See the ggplot2 cheat sheet to understand the ggplot syntax

p = ggplot(mRNA.Exp, aes(factor(gene), z.score))
p +
  geom_boxplot(aes(fill = factor(sample_type))) +
  # we drop some outlier, so plot looks nicer, this also causes the warning
  scale_y_continuous(limits = c(-1, 5)) +
  scale_fill_discrete(name = "Tissue")

Upon visual inspection do you see any genes that appear to be deferentially expressed in tumor tissue TP vs normal tissue NT?

Explore_FireBrowseR_BRCA

Here I use the `package::function()` structure

R packages needed for downloading and analyzing GDAC firehose cancer data

Package information:

Explore_FireBrowseR_BRCA

Here I use the package::function() structure

R packages needed for downloading and analyzing GDAC firehose cancer data

Package information:

Here I use the `package::function()` structure