workflow_hpv_patients.Rmd

---
title: "Light/Dark Pathways Workflow for HPV +/- Patients"
author: "Ted Laderas"
date: "6/6/2019"
output: html_notebook
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Installing the packageDir package

The majority of the analysis code is contained in a package called `packageDir` authored by Sam Higgins.

To install, use the following code:

```{r eval=FALSE}
install.packages("remotes")
remotes::install_github("biodev/packageDir")
```

## Reproducing HNSCC notebook

This notebook and repo contains everything needed to reproduce the light/dark pathway analysis in Choonoo, et al.

```{r}
library(here)
library("packageDir")
library("roxygen2")
library("DT")
library("dplyr")
```

## Processing Mutation Files

Before we do the analysis, we need to generate the two mutation files for our two different analyses: HPV positive patients, and HPV negative patients.

```{r eval=FALSE}
hpv_patients <- read.csv("data/HPV_Annotation_MB.csv")

#show breakdown of patients
table(hpv_patients$FINAL_HPV_CALL)

hpv_positive_patients <- hpv_patients %>% filter(FINAL_HPV_CALL == "Highest Confidence Positive") %>% pull(TCGA_BARCODE) %>% as.character()

hpv_negative_patients <- hpv_patients %>% filter(FINAL_HPV_CALL == "High Confidence Negative") %>% pull(TCGA_BARCODE) %>% as.character()
```

```{r eval=FALSE}
maf_file <- read.delim(here::here("data/TCGA.HNSC.mutect.84c7a87a-9dcc-48fb-bd69-ba9d6e6f3ca2.DR-7.0.somatic_cleaned.maf"))

maf_positive_patients <- maf_file %>% filter(TCGA_Patient_Barcode %in% hpv_positive_patients)

nrow(maf_positive_patients)
write.table(maf_positive_patients, here::here("data/TCGA.HNSC.HPV-positive.maf"), row.names=FALSE, sep="\t")

maf_negative_patients <- maf_file %>% filter(TCGA_Patient_Barcode %in% hpv_negative_patients)

nrow(maf_negative_patients)
write.table(maf_negative_patients, here::here("data/TCGA.HNSC.HPV-negative.maf"), row.names=FALSE, sep="\t")
```

## Processing Pathway Information

This step takes a little bit of time, so we've cached the results in `data/path_detail.rds`. Make `eval=TRUE` to run this part of the workflow.

```{r eval=FALSE}
path_detail <- getDefaultPaths("reference_data/paths/ReactomePathways 2015 02 17 13 46 25 2019.06.10 18.48.35.txt")
saveRDS(path_detail, file="data/path_detail.rds")
```

## Running Overlap analysis

A number of settings options have to be set for this analysis. 




### Somatic Mutation Settings for HPV Positive Analysis

```{r}
path_detail <- readRDS("data/path_detail.rds")

settings <- getBasicSettings()

#modify the settings object
settings$somatic_mutation_aberration_summary$interactive <- "n"

settings$somatic_mutation_aberration_summary$`Analyze pathways for individual members of the cohort? (y or n) `<- "n"

settings$somatic_mutation_aberration_summary$`Select a .maf file containing the data set to be analyzed.` <- "data/TCGA.HNSC.HPV-positive.maf"

settings$somatic_mutation_aberration_summary$`Would you like to include PolyPhen analysis results in this analysis? (y/n) `<- "n"

settings$somatic_mutation_aberration_summary$`Have manual gene symbol corrections already been conducted? (y/n)` <- "y"

settings$somatic_mutation_aberration_summary$`Analyze pathways for individual members of the cohort? (y or n) ` <- "n"

settings$somatic_mutation_aberration_summary$`Have manual gene symbol corrections already been conducted? (y/n)` <- "y"

settings$somatic_mutation_aberration_summary$`
Please enter the row numbers of the variant types you would like to analyze (sepparated by a space).
` <- "\nFrame_Shift_Del; Frame_Shift_Ins; In_Frame_Del; In_Frame_Ins; Missense_Mutation; Nonsense_Mutation; Nonstop_Mutation; Splice_Site\n"

settings$somatic_mutation_aberration_summary$mutation_type <- c("Frame_Shift_Del", "Frame_Shift_Ins", "In_Frame_Del", "In_Frame_Ins", "Missense_Mutation", "Nonsense_Mutation", "Nonstop_Mutation", "Splice_Site")


settings$somatic_mutation_aberration_summary$`Would you like to filter out hypermutators?
If yes, please enter a mutation count threshold.
If no just press enter n ` <- "n"

settings$somatic_mutation_aberration_summary$`Use special path significance analysis settings for this data type? (y/n)` <- "n"
 
```


# Process Functional Drug Screen Settings

```{r}
settings$functional_drug_screen_summary$`Please select a file with a drug screen results data set
` <- "data/Targetome_Level123_8_7_17.txt"

settings$functional_drug_screen_summary$interactive <- "n"

settings$functional_drug_screen_summary$`Have manual symbol corrections been performed yet for the current data set? (y/n)` <- "y"

settings$functional_drug_screen_summary$`Analyze pathways for individual members of the cohort? (y or n) ` <- "n"

settings$functional_drug_screen_summary$gene_stat <- "g"

settings$functional_drug_screen_summary$`
 To analyze drug screen panel coverage (for a panel that has or has not been run), enter p
 To process drug screen result set enter d
 To save an HTML summary of the results enter h
 To exit drug screen interface, enter q
` <- "p"
 
settings$functional_drug_screen_summary[['Enter \\"g\\" to examine coverage using a set of gene names.\nEnter \\"d\\" to examine coverage using drug names, along with a drug target matrix: ']] <- "g"

settings$functional_drug_screen_summary$`Please select a file` <- "data/Targetome_Level123_8_7_17.txt"

settings$functional_drug_screen_summary$`Please type in the name of the column with the gene symbols: ` <-	"targets"
settings$functional_drug_screen_summary$`Have manual gene symbol corrections already been made? (y/n)` <-	"y"

```

## Building the study with the settings

```{r}
study = getStudyObject(study.name="hnscc_hpv_positive", geneIdentifierType="HUGO", path_detail = path_detail, settings = settings)
```

## Running the Analysis

`loadBasicArms()` loads the data from the specified locations.

`autoRunFromSettings()` runs all of the mutation processing and overlap analysis code.

```{r}
study <- loadBasicArms(study)
study <- autoRunFromSettings(study)
saveStudy(study)

```

## Inspecting Output, HPV Positive Patients

Output is generated in `output/study_hnscc_hpv_positive/results/overlap_analysis`. The relevant files are:

`Aberrationally enriched, containing drug targets.txt` - Light pathway file

```{r}
lightPaths <- read.delim("output/study_hnscc_hpv_positive/results/overlap_analysis/Aberrationally enriched, containing drug targets.txt", sep="\t")
DT::datatable(lightPaths)

```

`Aberration enriched, not drug targeted.txt` - Dark pathway file

```{r}
darkPaths <- read.delim("output/study_hnscc_hpv_positive/results/overlap_analysis/Aberration enriched, not drug targeted.txt", sep = "\t")
DT::datatable(darkPaths)
```


# Running HPV negative patients

We can reuse the settings object, specifying our HPV negative patients mutation file:


```{r}
settings$somatic_mutation_aberration_summary$`Select a .maf file containing the data set to be analyzed.` <- "data/TCGA.HNSC.HPV-negative.maf"

study = getStudyObject(study.name="hnscc_hpv_negative", geneIdentifierType="HUGO", path_detail = path_detail, settings = settings)

study <- loadBasicArms(study)
study <- autoRunFromSettings(study)
saveStudy(study)
```

## Inspecting the output, HPV-Negative

Output is generated in `output/study_hnscc_hpv_negative/results/overlap_analysis`. The relevant files are:

`Aberrationally enriched, containing drug targets.txt` - Light pathway file

```{r}
lightPaths <- read.delim("output/study_hnscc_hpv_negative/results/overlap_analysis/Aberrationally enriched, containing drug targets.txt", sep="\t")
DT::datatable(lightPaths)

```

`Aberration enriched, not drug targeted.txt` - Dark pathway file

```{r}
darkPaths <- read.delim("output/study_hnscc_hpv_negative/results/overlap_analysis/Aberration enriched, not drug targeted.txt", sep = "\t")
DT::datatable(darkPaths)
```