derfinder (version 1.0.10)

mergeResults: Merge results from different chromosomes

Description

This function merges the results from running analyzeChr on several chromosomes and assigns genomic states using annotateRegions. It re-calculates the p-values and q-values using the pooled areas from the null regions from all chromosomes. Once the results have been merged, derfinderReport::generateReport can be used to generate an HTML report of the results. The derfinderReport package is available at https://github.com/lcolladotor/derfinderReport.

Usage

mergeResults(chrs = c(1:22, "X", "Y"), prefix = ".", significantCut = c(0.05, 0.1), genomicState, minoverlap = 20, mergePrep = FALSE, ...)

Arguments

chrs

The chromosomes of the files to be merged.

prefix

The main data directory path, which can be useful if analyzeChr is used for several parameters and the results are saved in different directories.

significantCut

A vector of length two specifiying the cutoffs used to determine significance. The first element is used to determine significance for the p-values and the second element is used for the q-values just like in calculatePvalues.

genomicState

A GRanges object created with makeGenomicState. It can be either the genomicState$fullGenome or genomicState$codingGenome component.

minoverlap

Determines the mininum overlap needed when annotating regions with annotateRegions.

mergePrep

If TRUE the output from preprocessCoverage is merged.

...

Arguments passed to other methods and/or advanced arguments.

Value

Seven Rdata files.

fullFstats.Rdata

Full F-statistics from all chromosomes in a list of Rle objects.

fullTime.Rdata

Timing information from all chromosomes.

fullNullSummary.Rdata

A DataFrame with the null region information: statistic, width, chromosome and permutation identifier. It's ordered by the statistics

fullRegions.Rdata

GRanges object with regions found and with full annotation from annotateNearest. Note that the column strand from annotateNearest is renamed to annoStrand to comply with GRanges specifications.

fullCoveragePrep.Rdata

A list with the pre-processed coverage data from all chromosomes.

fullAnnotatedRegions.Rdata

A list as constructed in annotateRegions with the assigned genomic states.

optionsMerge.Rdata

A list with the options used when merging the results. Used in derfinderReport::generateReport.

Details

If you want to calculate the FWER, supply optionsStats which #' is produced by analyzeChr.

See Also

analyzeChr, calculatePvalues, annotateRegions

Examples

## The output will be saved in the 'generateReport-example' directory
dir.create('generateReport-example', showWarnings = FALSE, recursive = TRUE)

## For convenience, the derfinder output has been pre-computed
file.copy(system.file(file.path('extdata', 'chr21'), package='derfinder',
mustWork=TRUE), 'generateReport-example', recursive=TRUE)

## Merge the results from the different chromosomes. In this case, there's
## only one: chr21
mergeResults(chrs='21', prefix='generateReport-example',
genomicState=genomicState$fullGenome)

## Not run: 
# ## You can then explore the wallclock time spent on each step
# load(file.path('generateReport-example', 'fullRegions.Rdata'))
# 
# ## Process the time info
# time <- lapply(fullTime, function(x) data.frame(diff(x)))
# time <- do.call(rbind, time)
# colnames(time) <- 'sec'
# time$sec <- as.integer(round(time$sec))
# time$min <- time$sec / 60
# time$chr <- paste0('chr', gsub('\\..*', '', rownames(time)))
# time$step <- gsub('.*\\.', '', rownames(time))
# rownames(time) <- seq_len(nrow(time))
# 
# ## Make plot
# library('ggplot2')
# ggplot(time, aes(x=step, y=min, colour=chr)) + geom_point() +
#     labs(title='Wallclock time by step') +
#     scale_colour_discrete(limits=chrs) +
#     scale_x_discrete(limits=names(fullTime[[1]])[-1]) + ylab('Time (min)') +
#     xlab('Step')
# ## End(Not run)