postanalysisloop: Batch analysis of samples in a QDNAseq-object for which...

Description Usage Arguments Details Value Note Author(s) See Also Examples

View source: R/ACE.R

Description

When models have been chosen for all (or just multiple) samples in a QDNAseq-object, this function can be used to perform a batch analysis on those samples. This encompasses printing segment data, printing copy number plots, and linking mutation data.

Usage

1
2
3
4
5
6
7
8
postanalysisloop(copyNumbersSegmented, modelsfile, variantdata, 
                 prefix="", postfix="", trncname=FALSE, inputdir=FALSE,
                 hetSNPs=FALSE, chrindex=1, posindex=2, freqindex,
                 altreadsindex, totalreadsindex, refreadsindex,
                 confidencelevel=FALSE, append=TRUE,
                 dontmatchnames=FALSE, printsegmentfiles=TRUE,
                 printnewplots=TRUE, imagetype='pdf',
                 outputdir="./", log=FALSE, segext='tsv')

Arguments

copyNumbersSegmented

QDNAseq-object with segmented data or file path of an rds-file containing a QDNAseq-object

modelsfile

Character string or data frame. When a character, it specifies the file path of a tab-delimited text containing model variables of samples. Expects columns with a header. It contains at least two columns: the first specifying the sample names and the second specifying the cellularity. The third column is the ploidy of the samples. When omitted, it is assumed to be 2. The fourth column is the standard of the samples. When omitted, it is calculated from the data in the object. The fitpicker.tsv file created by runACE can be used as modelsfile after cellularity of the likely fit is specified in the second column.

variantdata

Character string. Specifies directory containing variant data of samples. Optional. When argument inputdir is used, the function will first see if this argument is specified, if not it will check if the directory inputdir/variantdata exists, if not it will look for the variant files in inputdir itself. When inputdir is not used and this argument is omitted, the function will not link variant data. Mutation files need to have the same file extensions, which can be either .csv, .tsv, .txt, or .xls.

prefix

Character string. Used when a uniform character string precedes the sample name in the file name. E.g. "mutations_sample1.csv" has prefix "mutations_". Default = ""

postfix

Character string. As prefix, but then after the sample name. E.g. "sample1_somatics.csv" has postfix "_somatics". Default = ""

trncname

Logical. When TRUE, truncates the sample names of the QDNAseq-object starting from the first "_", or specify a character string with your regular expression of choice (trncname uses the gsub function). NOTE: use only when this will provide matches with the mutation files and the sample names in the modelsfile. Default = FALSE

inputdir

Character string. Specifies the directory which contains the files to be analyzed. Convenience function. Reduces the amount of arguments required when all data is available in the same directory: the QDNAseq-object, a file named "models.tsv" with the model parameters, and the mutation data (either in the inputdir itself or in a subdirectory "mutationdata"). Specifying the first arguments (copyNumbersSegmented, modelsfile, mutationdata) will take priority. When missing it will look in inputdir. When multiple rds-files are present in the inputdir, it will try the first one. Note: the path specified has no consequences for the location of the output. Default = FALSE

hetSNPs

Logical. If TRUE, half of the germline copies are assumed to be variant. Default = FALSE

chrindex

Integer. Column index in input file specifying the chromosome associated with the genomic location. Default = 1

posindex

Integer. Column index in input file specifying the position on the chromosome associated with the genomic location. Default = 2

freqindex

Integer. Column index in input file specifying the frequency (as a percentage) of the variant

altreadsindex

Integer. Column index in input file specifying the number of variant-supporting reads

totalreadsindex

Integer. Column index in input file specifying the read depth at the genomic location of the variant

refreadsindex

Integer. Column index in input file specifying the number of reference-supporting reads

confidencelevel

Numeric or logical. If read depth information is available, calculate the upper and lower bounds of this confidence level for the frequency and the number of variant copies of each variant. Will be skipped if FALSE. Default = FALSE

append

Logical. When TRUE, appends the output columns to the original mutation input file, but it still saves the result in a new file. When FALSE, the output file will only contain the columns "Chromosome", "Position", "Frequency", "Copynumbers", and "Mutant_copies" (and including the upper and lower bounds of the frequency and variant copies confidence interval, when applicable). Default = TRUE

dontmatchnames

Logical. When TRUE, the model variables are called by the index of the sample in the QDNAseq-object. This will only work if the order of samples in the object exactly matches the order of samples in the modelsfile. Use with caution! This is somewhat of an emergency option if for some reason the name matching is not working. I recommend trying to get the name matching to work. Default = FALSE

printsegmentfiles

Logical. When TRUE, prints a tab-delimited text file for each sample into a "segmentfiles" folder. Default = TRUE

printnewplots

Logical. When TRUE, prints plots into a "newplots" folder in the specified image type. Default = TRUE

imagetype

Character string specifying the image type graphics device. Default = "pdf"

outputdir

Character string. Save output into this custom directory. Default = "./"

log

Logical or integer. Use log conversion for creating segments output. Default = FALSE

segext

Character string specifying the extension for the segments output. Default = "tsv"

Details

If your input is tailored for this function, you could run it without any arguments! Most arguments help with matching sample names in the QDNAseq-object, the modelsfile, and the names and columns of the files containing variant data. You can "trim" the name of the file with variant data using the prefix (everything before the name) and postfix (everything after the name, but before the file extension) arguments to match your sample names. trncname migth help trimming the name in the QDNAseq-object, but be sure it still matches the sample names in the modelsfile (and mutation data file names when applicable).

Value

Prints the specified output to an indicated directory. Returns a list of copy number plots.

Note

The use of inputdir and outputdir should be fairly robust. However, using irregular file paths might cause problems. If you suspect problems with file paths, try setting the working directory to the intended inputdir.

Author(s)

Jos B. Poell

See Also

getadjustedsegments, linkvariants, runACE

Examples

1
2
3
4
5
6
7
## see the vignette for examples
## Not run: 
  data("copyNumbersSegmented")
  postanalysisloop(copyNumbersSegmented, "models.tsv", "variantdata", 
  outputdir = "loop_output")
  
## End(Not run)

ACE documentation built on Nov. 8, 2020, 5:30 p.m.