Assembling Clinical Information for the TCGA Ovarian Data
=========================================================

by Keith A. Baggerly

## 1 Executive Summary

### 1.1 Introduction

We want to produce an RData file with the
clinical information for the
ovarian cancer samples
profiled by TCGA.

### 1.2 Methods

We acquired the gzipped tarball containing the biotab
clinical information from the open access TCGA http
page,
[https://tcga-data.nci.nih.gov/tcgafiles/ftp_auth/distro_ftpusers/anonymous/tumor/ov/bcr/biotab/clin/](https://tcga-data.nci.nih.gov/tcgafiles/ftp_auth/distro_ftpusers/anonymous/tumor/ov/bcr/biotab/clin/),
on Sep 14, 2012.

We load the "clinical\_patient\_ov" information
into a data frame, and
construct an R "Surv" object for overall survival.

We also construct a binary indicator vector for the
presence or absence of residual disease (RD).

### 1.3 Results

We save
tcgaClinical,
tcgaOSYrs, and
tcgaRD
to the RData file "tcgaClinical.RData".

## 2 Libraries

We first load the libraries we will use
in this report.


```{r libraries, message=FALSE}

library(survival)

```


## 3 Specifying the Raw Data Location

Here, we specify the location of the data we acquired
from TCGA on our local system. You will need to acquire
these files and adjust this path before running
this report yourself.

```{r pathToTCGAData}

pathToTCGAData <-
    file.path("RawData","TCGA","Clinical")

```


## 4 Loading the Data

Here we simply load the table of clinical information.

```{r loadTCGAClinical}

tcgaClinical <-
    read.table(file.path(pathToTCGAData,"clinical_patient_ov.txt"),
               header=TRUE, sep="\t", row.names=1)
dim(tcgaClinical)
tcgaClinical[1,]

```

## 5 Defining Overall Survival

Next, we define an R "Surv" object for overall survival (OS).
We begin by looking at the recorded values for patient status.

```{r examineStatus}

table(tcgaClinical[,"vital_status"])
tcgaClinical[1:15,"days_to_death"]
tcgaClinical[1:15,"days_to_last_followup"]
tcgaClinical[1:15,"vital_status"]

```

Vital status is available for almost all of the 576 patients.
Checking the times available shows that days to death can
exceed the days to last followup (entry 1), and that days to
death is not available for patients still living, so we should
use the former for deceased patients and the latter for living
ones.

The above conclusions are based on a small sampling of
the data. We perform the more extensive sanity checks
for verification here.

```{r sanityChecks}

daysToDeath <-
    as.numeric(as.character(tcgaClinical[,"days_to_death"]))
daysToLastFollowup <-
    as.numeric(as.character(tcgaClinical[,"days_to_last_followup"]))
vitalStatus <-
    as.character(tcgaClinical[,"vital_status"])

summary(daysToDeath-daysToLastFollowup)
summary(daysToDeath[vitalStatus=="LIVING"])
daysToDeath[vitalStatus=="[Not Available]"]
daysToLastFollowup[vitalStatus=="[Not Available]"]

```

The sanity checks are passed.
We now assemble the Surv object.

```{r assembleSurv}

daysToEvent <- rep(NA, nrow(tcgaClinical))
daysToEvent[vitalStatus=="LIVING"] <-
    daysToLastFollowup[vitalStatus=="LIVING"]
daysToEvent[vitalStatus=="DECEASED"] <-
    daysToDeath[vitalStatus=="DECEASED"]
eventStatus <- rep(NA, nrow(tcgaClinical))
eventStatus[vitalStatus=="LIVING"] <- "Censored"
eventStatus[vitalStatus=="DECEASED"] <- "Uncensored"

tcgaOSYrs <- Surv(daysToEvent/365,eventStatus=="Uncensored")
rownames(tcgaOSYrs) <- rownames(tcgaClinical)

```


## 6 Defining a Residual Disease Indicator

Now we summarize the Residual Disease (RD) information
into a single indicator vector specifying if there
is any RD ("RD") or no RD ("No RD"). We begin
by tabulating the information we have.

```{r tableRDStatus}

table(tcgaClinical[,"tumor_residual_disease"])

```

We now define the indicator.

```{r specifyRDIndicator}

tcgaRD <- rep(NA,nrow(tcgaClinical))
tcgaRD[tcgaClinical[,"tumor_residual_disease"]==">20 mm"] <- "RD"
tcgaRD[tcgaClinical[,"tumor_residual_disease"]=="1-10 mm"] <- "RD"
tcgaRD[tcgaClinical[,"tumor_residual_disease"]=="11-20 mm"] <- "RD"
tcgaRD[tcgaClinical[,"tumor_residual_disease"]=="No Macroscopic disease"] <-
    "No RD"
table(tcgaRD)
names(tcgaRD) <- rownames(tcgaClinical)

```

## 7 Saving RData

Now we save the relevant information to an RData object.

```{r saveTCGAClinical}

save(tcgaClinical,
     tcgaOSYrs,
     tcgaRD,
     file=file.path("RDataObjects","tcgaClinical.RData"))

```


## 8 Appendix

### 8.1 File Location

```{r getLocation}

getwd()

```

### 8.2 Session Info 

```{r sessionInfo}

sessionInfo();

```