---
output: pdf_document
---

# Protocol to Build Interaction Network from PPI Data

Set knitr global properties
```{r setup, include=F}
knitr::opts_chunk$set(echo=T, message=F, warning=F)
```

## load common functions defined in Utils.R
```{r load Utils.R}
source("Utils.R")
```

## Package list needed
igraph: for function ecount()  
PSICQUIC: to fetch interaction data from public databases
```{r package list}
CRAN.packages <- c("igraph")
bioconductor.packages <- c("PSICQUIC")
```

## Package install and library loading
```{r libraries}
install.packages.if.necessary(CRAN.packages, bioconductor.packages)
```

## Set data directory, create it if necessary
```{r set data dir}
data.dir <- "networks"
if (!file.exists(data.dir)) {
  dir.create(data.dir)
}
```

## Protein-protein interaction Network from PSICQUIC + CCSB Interactome Database 
The goal is to construct a protein-protein interaction network containing binary direct physical protein interactions.  
Two sources are used :  
* The PSICQUIC portal, which allow to retrieve interaction data from different databases participating to the IMEX consortium. These databases describe the interactions in the PSI-MI format, allowing to select direct interactions between human proteins.  
Reference: del-Toro, N., Dumousseau, M., Orchard, S., Jimenez, R. C., Galeota, E., Launay, G., … Hermjakob, H. (2013). A new reference implementation of the PSICQUIC web service. Nucleic Acids Research, 41(Web Server issue), W601–6.  
* The Center for Cancer Systems Biology (CCSB) generates human interactome data. They have released in particular the dataset HI-II-2014 containing validated yeast 2-hybrid interactions between human proteins.  
Reference: Rolland, T., Taşan, M., Charloteaux, B., Pevzner, S. J., Zhong, Q., Sahni, N., … Vidal, M. (2014). A Proteome-Scale Map of the Human Interactome Network. Cell, 159(5), 1212–1226. doi:10.1016/j.cell.2014.10.050  
The code below creates networks from these 2 sources and merge them together

### Human PPI from PSICQUIC databases
Fetch human interactions (9096), from given list of databases, type direct (possibility to look for specific detection methods)
```{r fetch from PSICQUIC}
psicquic <- PSICQUIC()
# providers(psicquic) list all available DBs
DB.list <- c("BioGrid", "IntAct", "MatrixDB",
             "MBInfo", "MINT", "Reactome", "Reactome-FIs", "UniProt")
tbl.big <- PSICQUIC::interactions(psicquic,
                                  species="9606", 
                                  provider=DB.list,
                                # detectionMethod="experimental interaction detection",
                                  type="direct interaction")
```

Mapping gene names to HGNC  
Different gene name formats are used in the different databases. Everything is converted to HGNC gene symbol
```{r map names using biomart}
tbl.big <- addGeneInfo(IDMapper("9606"), tbl.big)
```

Extract columns of interest (protein A / protein B) and construct binary network  
Replace empty cells by NA  
Remove incomplete lines
```{r clean HGNC network}
colnames(tbl.big)[names(tbl.big) == "A.name"] <- "Symbol.A"
colnames(tbl.big)[names(tbl.big) == "B.name"] <- "Symbol.B"
NetHGNC <- clean.network(tbl.big)
```

### Human PPI from CCSB interactome database  
Download the data file from "http://interactome.dfci.harvard.edu/H_sapiens/index.php"  
```{r fetch CCSB data}
interactome.file <- "http://interactome.dfci.harvard.edu/H_sapiens/download/HI-II-14.tsv"
CCSB <- read.table(download.if.necessary(interactome.file, data.dir), header=T, sep="\t", as.is=T)
```

Select column of interest  
Replace empty cells by NA  
Remove incomplete lines
```{r clean CCSB network}
CCSB <- clean.network(CCSB)
```

### Merge the 2 PPI Networks
```{r merge}
net <- rbind (CCSB, NetHGNC)
```

### Graph conversion, simplification and export to file
```{r build igraph}
net <- build.network(as.matrix(net))
ecount(net)
gorder(net)
write.graph(net, file.path(data.dir, "PPI.gr"), format="ncol", weights=NULL)
```