%\VignetteIndexEntry{MetaGxOvarian: a package for ovarian cancer gene expression analysis}
%\VignetteDepends{xtable}
%\VignetteSuggests{}
%\VignetteKeywords{}
%\VignettePackage{MetaGxOvarian}

\documentclass[11pt]{article}

\usepackage[utf8]{inputenc}
\usepackage{authblk}
\usepackage{color}


\title{MetaGxOvarian: a package for ovarian cancer gene expression analysis}

\author[1,2]{Deena M.A. Gendoo}
\author[1]{Natchar Ratanasirigulchai}
\author[1]{Michael Zon}
\author[2]{Gregory Chen}
\author[3,4]{Levi Waldron}
\author[1,2]{Benjamin Haibe-Kains\thanks{benjamin.haibe.kains@utoronto.ca }}

\affil[1]{Bioinformatics and Computational Genomics Laboratory, Princess Margaret Cancer Center, University Health Network, Toronto, Ontario, Canada}
\affil[2]{Department of Medical Biophysics, University of Toronto, Toronto, Canada}
\affil[3]{Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Boston, MA, USA}
\affil[4]{Department of Biostatistics, Harvard School of Public Health, Boston, MA, USA}

\SweaveOpts{highlight=TRUE, tidy=TRUE, keep.space=TRUE, keep.blank.space=FALSE, keep.comment=TRUE}

<<setup,echo=FALSE,results=hide,eval=FALSE>>=
options(keep.source=TRUE, width = 50)
@

\begin{document}
\setkeys{Gin}{width=1.2\textwidth}
\SweaveOpts{concordance=TRUE}

\maketitle
\tableofcontents

%------------------------------------------------------------
\section{Installing the Package}
%------------------------------------------------------------

The MetaGxOvarian package is a compendium of Ovarian Cancer datasets.
The package is publicly available and can be installed from Bioconductor into R version 3.4.1 or higher.

To install the MetaGxOvarian package from Bioconductor:
<<install-pkg,eval=FALSE>>=
if (!requireNamespace("BiocManager", quietly=TRUE))
    install.packages("BiocManager")
BiocManager::install("MetaGxOvarian")
@

%------------------------------------------------------------
\section{Loading Datasets}
%------------------------------------------------------------
First we load the MetaGxOvarian package into the workspace.

To load the packages into R, please use the following commands:
<<loadlib>>=
library(MetaGxOvarian)
esets = MetaGxOvarian::loadOvarianEsets()[[1]]
@

This will load 26 expression datasets, with patients selected according to the default settings in the patientselection.config file. Users can modify the file to filter and annotate gene expression datasets and individual samples within them based on the following criteria:

\begin{description}
\item Datasets:  Conduct probe-gene mapping to select for the 'best' probe (default = TRUE)
\item Datasets:  Retain only genes that are common across all platforms loaded (default = FALSE)
\item Datasets:  Retain studies with a minimum sample size (default = 40)
\item Datasets:  Retain studies with a minimum umber of genes (default = 1000)
\item Datasets:  Retain studies with a minimum number of survival events
\item Datasets:  Remove duplicate samples (default = TRUE)
\item Datasets:  Rescale genes to Z-scores (default = FALSE)
\item Samples:  Ensure specific patient metadata is not missing
\item Samples:  Filter samples by sample type (tumour, healthy, etc)
\end{description}


%------------------------------------------------------------
\section{Obtaining Sample Counts in Datasets}
%------------------------------------------------------------

To obtain the number of samples per dataset, run the following:

<<sampleNumber-summary, results=tex>>=
numSamples <- vapply(seq_along(esets), FUN=function(i, esets){
  length(sampleNames(esets[[i]]))
  }, numeric(1), esets=esets)


SampleNumberSummaryAll <- data.frame(NumberOfSamples = numSamples,
                                     row.names = names(esets))
total <- sum(SampleNumberSummaryAll[,"NumberOfSamples"])
SampleNumberSummaryAll <- rbind(SampleNumberSummaryAll, total)
rownames(SampleNumberSummaryAll)[nrow(SampleNumberSummaryAll)] <- "Total"

require(xtable)
print(xtable(SampleNumberSummaryAll,digits = 2), floating = FALSE)
@
%------------------------------------------------------------
\section{Assess Phenotype Data}
%------------------------------------------------------------

We can also obtain a summary of the phenotype data (pData) for each expression dataset.
Here, we assess the proportion of samples in every datasets that contain a specific pData variable.
<<sampleNumberSummariesPdata, fig=TRUE>>=
#pData Variables
pDataID <- c("sample_type", "histological_type", "primarysite", "summarygrade",
             "summarystage", "tumorstage", "grade",
             "age_at_initial_pathologic_diagnosis", "pltx", "tax",
             "neo", "days_to_tumor_recurrence", "recurrence_status",
             "days_to_death", "vital_status")


pDataPercentSummaryTable <- NULL
pDataSummaryNumbersTable <- NULL

pDataSummaryNumbersList = lapply(esets, function(x)
  vapply(pDataID, function(y) sum(!is.na(pData(x)[,y])), numeric(1)))

pDataPercentSummaryList = lapply(esets, function(x)
  vapply(pDataID, function(y)
    sum(!is.na(pData(x)[,y]))/nrow(pData(x)), numeric(1))*100)

pDataSummaryNumbersTable = sapply(pDataSummaryNumbersList, function(x) x)
pDataPercentSummaryTable = sapply(pDataPercentSummaryList, function(x) x)

rownames(pDataSummaryNumbersTable) <- pDataID
rownames(pDataPercentSummaryTable) <- pDataID
colnames(pDataSummaryNumbersTable) <- names(esets)
colnames(pDataPercentSummaryTable) <- names(esets)

pDataSummaryNumbersTable <- rbind(pDataSummaryNumbersTable, total)
rownames(pDataSummaryNumbersTable)[nrow(pDataSummaryNumbersTable)] <- "Total"


# Generate a heatmap representation of the pData
pDataPercentSummaryTable<-t(pDataPercentSummaryTable)
pDataPercentSummaryTable<-cbind(Name=(rownames(pDataPercentSummaryTable))
                                ,pDataPercentSummaryTable)

nba<-pDataPercentSummaryTable
gradient_colors = c("#ffffff","#ffffd9","#edf8b1","#c7e9b4","#7fcdbb",
                    "#41b6c4","#1d91c0","#225ea8","#253494","#081d58")

library(lattice)
nbamat<-as.matrix(nba)
rownames(nbamat)<-nbamat[,1]
nbamat<-nbamat[,-1]
Interval<-as.numeric(c(10,20,30,40,50,60,70,80,90,100))

levelplot(nbamat,col.regions=gradient_colors,
          main="Available Clinical Annotation",
          scales=list(x=list(rot=90, cex=0.5),
                      y= list(cex=0.5),key=list(cex=0.2)),
          at=seq(from=0,to=100,length=10),
          cex=0.2, ylab="", xlab="", lattice.options=list(),
          colorkey=list(at=as.numeric(factor(c(seq(from=0, to=100, by=10)))),
                  labels=as.character(c( "0","10%","20%","30%", "40%","50%",
                                         "60%", "70%", "80%","90%", "100%"),
                                      cex=0.2,font=1,col="brown",height=1,
                                      width=1.4), col=(gradient_colors)))

@



%------------------------------------------------------------
\section{Session Info}
%------------------------------------------------------------
<<sessionInfo,echo=FALSE,results=tex,eval=TRUE>>=
toLatex(sessionInfo())
@

\end{document}