chicViewpointBackgroundModel

chicViewpointBackgroundModel computes for all given samples with all reference points a background model. For all relative distances to a reference point a negative binomial distribution is fitted. Moreover, for each relative distance to a reference point the average value for this location is computed. Both background models are used, the first one for p-value and significance computation, the second one to filter out interactions with a less x-fold over mean.

The background distributions are fixed at –fixateRange i.e. all distances lower / higher than this value use the fixed background distribution.

An example usage is:

$ chicViewpointBackgroundModel –matrices matrix1.cool matrix2.cool matrix3.cool –referencePoints referencePointsFile.bed –range 20000 40000 –outFileName background_model.bed

usage: chicViewpointBackgroundModel --matrices MATRICES [MATRICES ...]
                                    --referencePoints REFERENCEPOINTS
                                    [--averageContactBin AVERAGECONTACTBIN]
                                    [--outFileName OUTFILENAME]
                                    [--threads THREADS]
                                    [--fixateRange FIXATERANGE] [--help]
                                    [--version]

Required arguments

--matrices, -m

The input matrices (samples) to build the background model on.

--referencePoints, -rp

Bed file contains all reference points which should be used to build the background model.

Optional arguments

--averageContactBin

Average the contacts of n bins via a sliding window approach.

Default: 5

--outFileName, -o

The name of the background model file

Default: “background_model.bed”

--threads, -t

Number of threads. Using the python multiprocessing module.

Default: 4

--fixateRange, -fs

Fixate score of backgroundmodel starting at distance x. E.g. all values greater 500kb are set to the value of the 500kb bin.

Default: 500000

--version

show program’s version number and exit