RJMCMC: Nucleosome positioning mapping

In ArnaudDroitLab/RJMCMC: Bayesian hierarchical model for genome-wide profiling of nucleosome positions based on high-throughput short-read data (MNase-Seq data)

Description

Use of a fully Bayesian hierarchical model for chromosome-wide profiling of nucleosome positions based on high-throughput short-read data (MNase-Seq data). Beware that for a genome-wide profiling, each chromosome must be treated separatly.

Usage

rjmcmc(startPosForwardReads, startPosReverseReads, nbrIterations, kMax,
  lambda = 3, minInterval, maxInterval, minReads = 5,
  adaptIterationsToReads = TRUE)

Arguments

startPosForwardReads	a vector of numeric, the start position of all the forward reads.
startPosReverseReads	a vector of numeric, the start position of all the reverse reads. Beware that the start position of a reverse read is always higher that the end positition.
nbrIterations	a positive integer or numeric, the number of iterations. Non-integer values of nbrIterations will be casted to integer and truncated towards zero.
kMax	a positive integer or numeric, the maximum number of nucleosomes per region. Non-integer values of kMax will be casted to integer and truncated towards zero.
lambda	a positive numeric, the theorical mean of the Poisson distribution. Default: 3.
minInterval	a numeric, the minimum distance between two nucleosomes.
maxInterval	a numeric, the maximum distance between two nucleosomes.
minReads	a positive integer or numeric, the minimum number of reads in a potential canditate region. Non-integer values of minReads will be casted to integer and truncated towards zero. Default: 5.
adaptIterationsToReads	a logical indicating if the number of iterations must be modified in function of the number of reads. Default: TRUE.

Value

a list of class "rjmcmcNucleosomes" containing:

• call the matched call.

• K a vector of integer, the estimation of the number of the nucleosomes for each iteration.

• k a integer, the final estimation of the number of nucleosomes.

• mu a vector of numeric of length k, the positions of the nucleosomes.

• sigmaf a vector of numeric of length k, the variance of the forward reads for each nucleosome.

• sigmar a vector of numeric of length k, the variance of the reverse reads for each nucleosome.

• delta a vector of numeric of length k, the distance between the maxima of the forward and reverse reads position densities for each nucleosome.

• df a vector of numeric of length k, the degrees of freedom for each nucleosome.

• w a vector of positive numerical of length k, the weight for each nucleosome. The sum of all w values must be equal to 1.

• qmu a matrix of numerical with a number of rows of k, the 2.5% and 97.5% quantiles of each mu.

• qsigmaf a matrix of numerical with a number of rows of k, the 2.5% and 97.5% quantiles of the variance of the forward reads for each nucleosome.

• qsigmar a matrix of numerical with a number of rows of k, the 2.5% and 97.5% quantiles of the variance the reverse reads for each nucleosome.

• qdelta a matrix of numerical with a number of rows of k, the 2.5% and 97.5% quantiles of the distance between the maxima of the forward and reverse reads position densities for each nucleosome.

• qdf a matrix of numerical with a number of rows of k, the 2.5% and 97.5% quantiles of the degrees of freedom for each nucleosome.

• qw a matrix of numerical with a number of rows of k, the 2.5% and 97.5% quantiles of the weight for each nucleosome.

Author(s)

Rawane Samb, Pascal Belleau, Astrid Deschênes

Examples

## Loading dataset
data(reads_demo)

## Nucleosome positioning, running both merge and split functions
result <- rjmcmc(startPosForwardReads = reads_demo$readsForward,
         startPosReverseReads = reads_demo$readsReverse,
         nbrIterations = 1000, lambda = 2, kMax = 30,
         minInterval = 146, maxInterval = 292, minReads = 5)

## Print the final estimation of the number of nucleosomes
result$k

## Print the position of nucleosomes
result$mu

ArnaudDroitLab/RJMCMC documentation built on May 5, 2019, 7:06 a.m.