---
title: "Creating Target Decoy Databases"
author: "Witek Wolski"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Creating Target Decoy Databases}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---



Using the package prozor for creating standardized fasta.
The package can be used to: 
     
     * merge several fasta files into a single fasta file
     * add reverse sequences to the fasta file.
     * add contaminants to the fasta file


# Contaminants

This package provides two sets of typical contaminant proteins and peptides, one with and one without contaminants of human origin, which can be accessed by the functions `loadContaminantsFasta` and `loadContaminantsNoHumanFasta`.
At the [FGCZ](https://fgcz.ch/) we always are adding one of those two contaminant files to the database. To databases already containing human proteins, we will add the `ContaminantsNoHumanFasta`.
The contaminants are easy to distinguish from other entries thanks to the `zz|FGCZCont` prefix.

```{r}
library(prozor)
head(names(loadContaminantsFasta2021()))
length(loadContaminantsFasta2021())
length(loadContaminantsFasta2021(noHuman = TRUE))

```

# Creating a fasta protein amino acid sequence database for searching.

To merge several _fasta_ databases into a single file place them into a single folder and give the folder the name of the database.
At the [FGCZ](https://fgcz.ch/) the database name starts with the project number e.g. `p1000` a consecutive number e.g. `db1` and descriptive name `example,i.e. `p1000_db1_example`.

Add to the folder also an annotation.txt file. The annotation file should contain a single line formatted like a _fasta_ entry header with the following conent:
`aa|p<project_number>_<database_name>|<YYYYMMDD> <detailed_description>`.

Example : `AA|p1000_db1_example|20180119_Example https://github.com/protViz/prozor`


The package provides an example of such a folder with the fasta files. Based on this folder a database can be created.

```{r}
databasedirectory = system.file("p1000_db1_example",package = "prozor")
#databasedirectory <- file.path(find.package("prozor"), "p1000_db1_example")
dbname <- basename(databasedirectory)
fasta <- grep("fasta", dir(databasedirectory),value = TRUE)
files1 <- file.path(databasedirectory,fasta)
annot <- grep("annotation",dir(databasedirectory), value = TRUE)
annotation <- readLines(file.path(databasedirectory,annot))
annotation

```

## Create non decoy database


```{r}
resDB <- createDecoyDB(files1, useContaminants = loadContaminantsFasta2021(),
                       annot = annotation, revLab = NULL)
length(resDB)

```

Based on the directory name we build the name of the fasta file adding the current date.

```{r}
dirname(databasedirectory)
xx <- file.path(dirname(databasedirectory), paste(dbname,"_",format(Sys.time(), "%Y%m%d"),".fasta" ,sep = ""))
print(xx)

```


```{r eval=FALSE}
writeFasta(resDB, file=xx)

```

## Create decoy database

To add a decoy database, using reverse sequences specify the `revLab` parameter in the `createDecoyDB` function.
The resulting database will be twice as long as the non-decoy database.


```{r}
resDBDecoy <- createDecoyDB(files1,
                            useContaminants = loadContaminantsFasta2021(),
                            annot = annotation,
                            revLab = "REV_")

resDBDecoy[[length(resDBDecoy) - 1]]
length(resDBDecoy)

sum(duplicated(names(resDBDecoy)))
sum(duplicated(resDBDecoy))

dbname_decoy <- unlist(strsplit(dbname,"_"))
dbname_decoy <- paste(c(dbname_decoy[1],"d",dbname_decoy[2:length(dbname_decoy)]),collapse = "_")
dbname_decoy

xx <- file.path(dirname(databasedirectory), paste(dbname_decoy,"_",format(Sys.time(), "%Y%m%d"),".fasta" ,sep = ""))
print(xx)

```


```{r eval=FALSE}
writeFasta(resDBDecoy, file = xx)
```


# Session Info

```{r}
sessionInfo()
```