runCHR: Run 'rjmcmc' on multiples segments and merge results.

In ArnaudDroitLab/RJMCMCNucleosomes: Bayesian hierarchical model for genome-wide nucleosome positioning with high-throughput short-read data (MNase-Seq)

Description

Run rjmcmc on a segment that is contained in a list of segments. Files generated by the function are all saved in a directory specified by user. A RData log file is created when a segment has been run while the result is saved in a RDS file.

If the same output directory is used more than once, the rjmcmc won't be called for segments that have the <c3><a0> corresponding RData log file.

Usage

runCHR(p, seg, niter, kmax, lambda, ecartmin, ecartmax, minReads,
  adaptNbrIterations, vSeed = -1, saveAsRDS = FALSE, dirOut = "out")

Arguments

p	a integer, the position of the segment to treat in a list of GRanges.
seg	a list a GRanges containing the segments to be process.
niter	a positive integer or numeric, the number of iterations. Non-integer values of nbrIterations will be casted to integer and truncated towards zero.
lambda	a positive numeric, the theorical mean of the Poisson distribution. Default: 3.
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.
adaptNbrIterations	a logical indicating if the number of iterations must be modified in function of the number of reads.
vSeed	a integer. A seed used when reproducible results are needed. When a value inferior or equal to zero is given, a random integer is used. Default: -1.
saveAsRDS	a logical. When TRUE, a RDS file containing the complete output of the c++ rjmcmc() function is created. Default : FALSE.
kMax	a positive integer or numeric, the maximum number of degrees of freedom per region. Non-integer values of kMax will be casted to integer and truncated towards zero.
minInterval	a numeric, the minimum distance between two nucleosomes.
maxInterval	a numeric, the maximum distance between two nucleosomes.
maxLength	a positive integer or numeric, the length of each segment.

Value

0.

Author(s)

Pascal Belleau, Astrid Deschenes

Examples

## Load synthetic dataset of reads
data(syntheticNucleosomeReads)

## Use dataset of reads to create GRanges object
sampleGRanges <- GRanges(seqnames = syntheticNucleosomeReads$dataIP$chr,
    ranges = IRanges(start = syntheticNucleosomeReads$dataIP$start,
    end = syntheticNucleosomeReads$dataIP$end),
    strand = syntheticNucleosomeReads$dataIP$strand)

# Segmentation of the reads
seg <- segmentation(sampleGRanges, zeta = 147, delta = 50, maxLength = 1000)
## Not run: 
dir.create("out")
dir.create("out/done")
dir.create("out/results")

runCHR(p=1, seg=seg, niter=1000, kmax=330,lambda=3,
            ecartmin=147, ecartmax=297, minReads=5)
## End(Not run)

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