GenoGAMSettings-class: GenoGAMSettings
In gstricker/fastGenoGAM: A GAM based framework for analysis of ChIP-Seq data

Description Usage Arguments Details Value Slots Author(s) Examples

This class is designed to store global settings for the computation of the GenoGAM package

1	GenoGAMSettings(...)

...

Any parameters corresponding to the slots and their possible values.

Center can have three values: TRUE, FALSE, NULL. TRUE will trigger the center function, FALSE will trigger the use of the entire fragment. NULL should be used in case a custom process function is used. In case a custom function is used, it has to satisfy the following: It has to handle a GAlignments object as input and output a GRanges object of regions, e.g. fragments. This regions are in turn used to compute the coverage via the IRanges::coverage function. Note, that there is a difference between the GAlignments object in the single and paired end case.

optimControl has two maxit fields: 'maxit' refers to the maximal iterations in the cross validation procedure. Convergence rarely exceeds 50. 'betaMaxit' refers to the maximum iterations in estimation of the beta parameters. Convergence will sometimes need up to 1000 iterations. Also the parameter 'trace' is overwritten. Please use the threshold setting through futile.logger::flog() to control trace information. All other parameters can be used as specified in ?optim. Note that the method used to estimate the beta vector is 'L-BFGS-B'. The method for cross validation can be changed. Keep in mind however, that not all methods will work, as it is not gradient based.

An object of class GenoGAMSettings

center: A logical or NULL value to specify if the raw data should be centered, i.e. only the midpoint of the fragment will be used to represent its coverage. See details.
chromosomeList: A character vector of chromosomes to be used. NULL for all chromosomes.
bamParams: An object of class ScanBamParam. See ?Rsamtools::ScanBamParam for possible settings. Usually used to set specific ranges, to read in.
processFunction: A custom function on how to process raw data. Not used if center is TRUE/FALSE. This is not intended for the user, but if needed anyway, see details.
optimMethod: The optiomisation method to be used in cross validation. See ?optim for a complete list.
optimControl: List of control settings for the optim function. Almost all parameters are supported, with a couple of exceptions. See details. For a complete list of parameters see ?optim.
estimControl: List of control settings for the parameter estimation algorithm.
hdf5Control: List of control settings for the HDF5 backend
dataControl: List of control settings for the processed data. The size of the region to use for the computation of the count matrix, that is later used by DESeq2. Also the size of the regions that will be used for Cross Validation. And the spacing between knots.

Georg Stricker georg.stricker@in.tum.de

# Construct the class
GenoGAMSettings()

# Construct the class with custom parameters
## specify chromosomes
center <- FALSE
chromosomeList <- c('chr1', 'chr2')
GenoGAMSettings(center = center, chromosomeList = chromosomeList)

## Specify ranges
gr <- GenomicRanges::GRanges("chr1", IRanges(1, 10000))
bamParams <- Rsamtools::ScanBamParam(which = gr)
GenoGAMSettings(bamParams = bamParams, center = TRUE)