---
title: "Generate synthetic nucleosome maps"
output:
  BiocStyle::html_document:
    toc: true
vignette: >
  %\VignetteIndexEntry{Generate synthetic nucleosome maps}
  %\VignettePackage{nucleoSim}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r style, echo = FALSE, results = 'asis', message = FALSE}
BiocStyle::markdown()
library(knitr)
```

<br>
**Package**: `r Rpackage("nucleoSim")`<br />
**Authors**: `r packageDescription("nucleoSim")[["Author"]]`<br />
**Version**: `r packageDescription("nucleoSim")$Version`<br />
**Compiled date**: `r Sys.Date()`<br />
**License**: `r packageDescription("nucleoSim")[["License"]]`<br />

# Licensing and citing

This package and the underlying `r Rpackage("nucleoSim")` code are 
distributed under the Artistic license 2.0. You are free to use and 
redistribute this software.

If you use this package for a publication, we would ask you to cite the 
following:

> Samb R, Khadraoui K, Belleau P, et al. (2015). "Using informative Multinomial-Dirichlet prior in a t-mixture with reversible jump estimation of nucleosome positions for genome-wide profiling." Statistical Applications in Genetics and Molecular Biology. Volume 14, Issue 6, Pages 517–532, ISSN (Online) 1544-6115, ISSN (Print) 2194-6302, December 2015. <a href="http://dx.doi.org/10.1515/sagmb-2014-0098">doi: 10.1515/sagmb-2014-0098</a>

> Flores O et Orozco M (2011). “nucleR: a package for non-parametric nucleosome positioning.” Bioinformatics, 27, pp. 2149–2150. <a href="http://dx.doi.org/10.1093/bioinformatics/btr345">doi: 10.1093/bioinformatics/btr345</a>


# Introduction

`r Rpackage("nucleoSim")` can simulate datasets for nucleosomes experiments. 

The `r Rpackage("nucleoSim")` package generates synthetic maps with 
sequences covering nucleosome regions as well as synthetic maps with 
forward and reverse reads (paired-end reads) emulating next-generation 
sequencing. Furthermore, synthetic hybridization data of "Tiling Arrays" can 
also be generated.

The `r Rpackage("nucleoSim")` package allows the user to introduce various 
'contaminants' into the sequence datasets, such as fuzzy nucleosomes and 
missing nucleosomes, in order to be more realistic and to enable the 
evaluation of the influence of common 'noise' on the detection of nucleosomes.

* Select the number of well-positioned nucleosomes to generate
* Select the number of well-positioned nucleosomes to randomly delete
* Select the number of fuzzy nucleosomes to  add
* Select the type of distribution used to assign the 
start position to the sequences associated with the nucleosomes
* Select the variance associated with the distribution of the start position 
of the sequences for the well-positioned nucleosomes (or the start position of
the forward reads for paired-end reads)
* Select the variance associated with the distribution of the start position 
of the sequences for the fuzzy nucleosomes (or the start position of
the forward reads for paired-end reads)
* Select the variance associated with the Normal distribution used to assign
the length of each sequence (or the distance between start positions of 
paired-end reads)
* Select the maximum sequence coverage associated with one nucleosome
* Select the length of the DNA linker
* Select the length of the nucleosomes (not recommended changing default value)
* Select a seed when results need to be reproducible

The `r Rpackage("nucleoSim")` package has been largely inspired by the
*Generating synthetic maps* section of the 
Bioconductor `r Biocpkg("nucleR")` package (Flores et Orozco, 2011).


# Loading `r Rpackage("nucleoSim")` package

As with any R package, the `r Rpackage("nucleoSim")` package should first be 
loaded with the following command:

```{r loadingPackage, warning=FALSE, message=FALSE}
library(nucleoSim)
```


# Description

The packages can generate 2 types of synthetic data sets:

**Synthetic Nucleosome Maps**: A map with complete sequences covering the 
nucleosome regions
    
**Synthetic Nucleosome Samples**: A map with forward and reverse reads 
(paired-end reads) emulating those obtained using a next-generation 
sequencing technology on a nucleosome map 


## Synthetic Nucleosome Maps

A synthetic nucleosome map is a section of genome covered by a fixed number
of nucleosomes. Each nucleosome being associated with a specific number of 
sequences. The parameters passed to the `syntheticNucMapFromDist()` function are
going to affect the distribution of the nucleosomes, as well as, the sequences
associated with each nucleosome.

Technically, the synthetic nucleosome map is separated into 3 steps:

**1. Adding well-positioned nucleosomes**

The synthetic nucleosome map is split into a fixed number of 
sections (`wp.num`) of  equal length (`(nuc.len + lin.len)` 
bases). The center of the nucleosomes is positioned at a fixed number of bases 
from the beginning of each section (`round(nuc.len/2)` bases). Sequences are 
assigned, to each nucleosome, using an uniform distribution. The number of 
sequences, assigned to each nucleosome, can vary from 1 to `max.cover`. 
The distribution (`distr`), as 
well as the variance (`wp.var`) are used to add some fluctuation to the 
starting position of the sequences, which as a mean position corresponding to 
the starting position of a region. Some fluctuation of the length of 
the sequence is also added following
a normal distribution with a fixed variance (`len.var`). The mean 
length of the sequences corresponds to the length of the 
nucleosomes (`nuc.len`).

**2. Deleting some well-positioned nucleosomes**

A fixed number of nucleosomes (`wp.del`) are deleted. Each nucleosome has 
an equal probability to be deleted.  A 
nucleosome is considered deleted when all sequences associated 
with it are eliminated.

**3. Adding fuzzy nucleosomes**

A fixed number of fuzzy nucleosomes (`fuz.num`) are added. The position of the
fuzzy nucleosomes is selected following an uniform distribution. Such as for
the well-positioned nucleosomes, sequences are 
assigned, to each fuzzy nucleosome, using an uniform distribution. The number 
of sequences assigned can vary from 1 to `max.cover`. 
The distribution (`distr`), as 
well as the variance (`wp.var`) are used to add some fluctuation to the 
starting position of the sequences, which as a mean position corresponding to 
the starting position of a region. Some fluctuation of the length of 
the sequence is also added following
a normal distribution with a fixed variance (`len.var`). The mean 
length of the sequences corresponds to the length of the 
nucleosomes (`nuc.len`).

### Create a nucleosome map using syntheticNucMapFromDist()

This is an example showing how a synthetic nucleosome map can be generated.

```{r demoMap, warning=FALSE, message=FALSE, collapse=TRUE}
wp.num           <- 20         ### Number of well-positioned nucleosomes
wp.del           <- 5          ### Number of well-positioned nucleosomes to delete
wp.var           <- 30         ### variance associated with the starting 
                               ###   position of the sequences of the 
                               ###   well-positioned nucleosomes
fuz.num          <- 5          ### Number of fuzzy nucleosomes
fuz.var          <- 50         ### Variance associated with the starting 
                               ###   positions of the sequences for the 
                               ###   fuzzy nucleosomes
max.cover        <- 70         ### Maximum sequences associated with one 
                               ###   nucleosome (default: 100)
nuc.len          <- 147        ### Length of the nucleosome (default: 147)
len.var          <- 12         ### variance associated with the length of 
                               ###   the sequences (default: 10)
lin.len          <- 20         ### Length of the DNA linker (default: 20)
distr            <- "Normal"   ### Type of distribution to use

rnd.seed         <- 210001     ### Set seed when result needs to be reproducible
        
#### Create a synthetic nucleosome map
nucleosomeMap <- syntheticNucMapFromDist(wp.num=wp.num, wp.del=wp.del, wp.var=wp.var, 
                                    fuz.num=fuz.num, fuz.var=fuz.var, max.cover=max.cover,
                                    nuc.len=nuc.len, len.var=len.var, lin.len=lin.len, 
                                    rnd.seed=rnd.seed, distr=distr)

#### The start positions of all well-positioned nucleosomes
nucleosomeMap$wp.starts

#### The number of sequences associated with each well-positioned nucleosome
nucleosomeMap$wp.nreads

#### A IRanges object containing all sequences for the well-positioned nucleosomes
head(nucleosomeMap$wp.reads, n = 2)

#### The start positions of all fuzzy nucleosomes
nucleosomeMap$fuz.starts

#### The number of sequences associated with each fuzzy nucleosome
nucleosomeMap$fuz.nreads

#### A IRanges object containing all sequences for the fuzzy nucleosomes
head(nucleosomeMap$fuz.reads, n = 2)

#### A IRanges object containing all sequences for all nucleosomes
head(nucleosomeMap$syn.reads, n = 2)
```

The synthetic nucleosome map can easily be visualized using `plot()` function.
On the graph, each nucleosome is located on the graph using the 
coordonnates:

`   (x,y) = 
(the central position of the nucleosome, the number of sequences 
associated with the nucleosome)`

```{r showMap, fig.align='center', fig.height=5, fig.width=8}
#### Create visual representation of the synthetic nucleosome map
plot(nucleosomeMap, xlab="Position", ylab="Coverage")
```

### Simulate hybridization data of *Tiling Arrays*

The `syntheticNucMapFromDist()` function contains an option (`as.ratio`) which
enable the simulation of hybridization data of "*Tiling Arrays*". The data are
generated by calculating the ratio between the nucleosome map and 
a control map of random sequences created using a uniform distribution. The
control map simulates a DNA randomly fragmented sample.

This is an example showing how a synthetic nucleosome map can be generated.

```{r demoMapTiling, warning=FALSE, message=FALSE, collapse=TRUE}
as.ratio         <- TRUE       ### Activate the simulation of hybridization data
rnd.seed         <- 212309     ### Set seed when result needs to be reproducible
        
#### Create a synthetic nucleosome map with hybridization data
nucleosomeMapTiling <- syntheticNucMapFromDist(wp.num=10, wp.del=2, wp.var=20, 
                                    fuz.num=1, fuz.var=32, max.cover=50,
                                    nuc.len=145, len.var=3, lin.len=40, 
                                    rnd.seed=rnd.seed, as.ratio=as.ratio,
                                    distr="Uniform")

#### Control sequences for hybridization data (only available when as.ratio is TRUE)
head(nucleosomeMapTiling$ctr.reads, n=4)

#### Ratio for hybridization data (only available when as.ratio is TRUE)
head(nucleosomeMapTiling$syn.ratio, n=4)

#### Create visual representation of the synthetic nucleosome map
plot(nucleosomeMapTiling)
```


## Synthetic Nucleosome Samples

A synthetic nucleosome sample is a map with forward and reverse reads 
(paired-end reads) emulating those obtained using a next-generation 
sequencing technology. It is created using the same first 3 steps than the 
synthetic nucleosome map. However, some new steps are present:

**1. Adding well-positioned nucleosomes **

The synthetic nucleosome map is split into a fixed number of 
sections (`wp.num`) of  equal length (`(nuc.len + lin.len)` 
bases). The center of the nucleosomes are positioned at a fixed number of 
bases from the beginning of each section (`round(nuc.len/2)` bases). 
Paired-end reads are assigned, to each nucleosome, using an uniform 
distribution. The number of paired-end reads, assigned to each 
nucleosome, can vary from 1 to `max.cover`. The distribution (`distr`), as 
well as the variance (`wp.var`) are used to add some fluctuation to the 
starting position of the forward reads, which as a mean position corresponding 
to the starting position of a region. Some fluctuation of the distance between 
start positions of paired-end reads is added following
a normal distribution with a fixed variance (`len.var`). The mean 
distance between start positions of paired-end reads corresponds to 
the length of the nucleosomes (`nuc.len`).

**2. Deleting some well-positioned nucleosomes**

A fixed number of nucleosomes (`wp.del`) are deleted. Each nucleosome has 
an equal probability to be deleted. A nucleosome is considered deleted 
when all paired-end reads 
associated with it are eliminated.

**3. Adding fuzzy nucleosomes**

A fixed number of fuzzy nucleosomes (`fuz.num`) are added. The position of the
fuzzy nucleosomes is selected following an uniform distribution. Such as for
the well-positioned nucleosomes, reads are 
assigned, to each fuzzy nucleosome, using an uniform distribution. The number 
of paired-end reads assigned can vary from 1 to `max.cover`. 
The distribution (`distr`), as 
well as the variance (`wp.var`) are used to add some fluctuation to the 
starting position of the forward reads, which as a mean position corresponding 
to the starting position of a region. Some fluctuation of the distance between 
start positions of paired-end reads is also added following
a normal distribution with a fixed variance (`len.var`). The mean distance 
between start positions of paired-end reads corresponds to the length of 
the nucleosomes (`nuc.len`). All reads have a fixed length
(`read.len`).

**4. Adding an offset**

An offset (`offset`) is added to all nucleosomes and
reads positions to ensure that all values are positive (mainly pertinent for
reads).


### Create a nucleosome sample using syntheticNucReadsFromDist()

This function needs information about the nucleosomes and their distribution 
to generate a nucleosome sample. The output is of class `syntheticNucMap`.

```{r demoSample, warning=FALSE, message=FALSE, collapse=TRUE}
wp.num           <- 30            ### Number of well-positioned nucleosomes
wp.del           <- 10            ### Number of well-positioned nucleosomes to delete
wp.var           <- 30            ### variance associated with the starting 
                                  ###   positions of the sequences for the 
                                  ###   well-positioned nucleosomes
fuz.num          <- 10            ### Number of fuzzy nucleosomes
fuz.var          <- 50            ### Variance associated with the starting 
                                  ###   positions of the sequences for the 
                                  ###   fuzzy nucleosomes
max.cover        <- 90            ### Maximum paired-end reads associated with 
                                  ###   one nucleosome (default: 100)
nuc.len          <- 147           ### Length of the nucleosome (default: 147)
len.var          <- 12            ### variance associated with the distance 
                                  ###   between start positions of 
                                  ###   paired-end reads (default: 10)
lin.len          <- 20            ### Length of the DNA linker (default: 20)
read.len         <- 45            ### Length of the generated forward and 
                                  ###   reverse reads (default: 40)
distr            <- "Uniform"     ### Type of distribution to use
offset           <- 10000         ### The number of bases used to offset 
                                  ###   all nucleosomes and reads

rnd.seed         <- 202           ### Set seed when result needs to be reproducible
        
#### Create Uniform sample
nucleosomeSample <- syntheticNucReadsFromDist(wp.num=wp.num, wp.del=wp.del, wp.var=wp.var,
                                        fuz.num=fuz.num, fuz.var=fuz.var, max.cover=max.cover,
                                        nuc.len=nuc.len, len.var=len.var, read.len=read.len, 
                                        lin.len=lin.len, rnd.seed=rnd.seed, distr=distr, 
                                        offset=offset)

#### The central position of all well-positioned nucleosomes with the
#### number of paired-end reads each associated with each one
head(nucleosomeSample$wp, n = 2)

#### The central position of all fuzzy nucleosomes with the
#### number of paired-end reads each associated with each one
head(nucleosomeSample$fuz, n = 2)

#### A data.frame with the name of the synthetic chromosome, the starting 
#### position, the ending position and the direction of all forward and
#### reverse reads
head(nucleosomeSample$dataIP, n = 2)
```

The synthetic nucleosome sample can easily be visualized using `plot()` 
function. On the graph, each nucleosome is located on the graph using the 
coordinates:

`   (x,y) = 
(the central position of the nucleosome, the number of paired-end reads associated with
the nucleosome)`

```{r showSample, fig.align='center', fig.height=5, fig.width=8}
#### Create visual representation of the synthetic nucleosome sample
plot(nucleosomeSample, xlab="Position", ylab="Coverage (number of reads)")
```

### Create a nucleosome sample using syntheticNucReadsFromMap()

A synthetic nucleosome sample can be created using a nucleosome map. The 
nucleosomes and reads present in the nucleosome map will be added an offset. 
Forward and reverse reads will also be generated. The output is of class
`syntheticNucMap`.

```{r demoSampleFromMap, warning=FALSE, message=FALSE, collapse=TRUE}

#### A nucleosome map has already been created 
class(nucleosomeMap)

#### 
read.len    <- 45   ### The length of the reverse and forward reads
offset      <- 500  ### The number of bases used to offset all nucleosomes and reads

#### Create nucleosome sample
nucleosomeSampleFromMap <- syntheticNucReadsFromMap(nucleosomeMap, 
                                        read.len=read.len, offset=offset)

#### A data.frame with the name of the synthetic chromosome, the starting 
#### position, the ending position and the direction of all forward and
#### reverse reads
head(nucleosomeSampleFromMap$dataIP, n = 2)
```


# Session info

Here is the output of `sessionInfo()` on the system on which this document was 
compiled:

```{r sessionInfo, echo=FALSE}
sessionInfo()
```