generate.binding.profile: Construct the binding profile ('kernel') required as input...

In MeDiChI: MeDiChI ChIP-chip deconvolution library

Description

Construct a model-based binding profile as described in Reference (1).

Usage

generate.binding.profile(fragment.distrib = function(x, ...) dgamma(x,
shape = 6, scale = 50), bs.size = 1, tile.size = 50, min.frag.size = 0,
positions = seq(0, 1001, by = 50), intensity.scaling = function(x, ...)
x, hybridization.prob = function(x, ...) as.integer(x > 10), interp = T,
plot = F, verbose = F, no.multicore=T, ...)

Arguments

fragment.distrib	Distribution of DNA fragment lengths.
bs.size	Footprint of TF binding site on the genome.
tile.size	Length of the array probes (in bp).
min.frag.size	Cutoff for minimum DNA fragment size.
positions	The distances from the center of the profile for which to compute the relative intensity. A longer vector increases the running time of this procedure. Values for distances other than those listed here are computed via interpolation. See 'interp'.
intensity.scaling	Intensity of DNA fragments as a function of their length.
hybridization.prob	Hybridization probability as a function of length of complementary sequence.
interp	If TRUE, interpolate the values for distances other than those listed in 'positions'.
plot	If TRUE, plot the resulting profile.
verbose	If TRUE, be verbose.
no.multicore	Prevent use of multiple cores, even if 'multicore' is installed.

...	Further parameters for 'fragment.distrib', 'intensity.scaling', and 'hybridization.prob'.

Details

No details.

Value

A two-column matrix containing positions (distance from center, in bp) and relative intensities of the profile.

Author(s)

David J Reiss, Institute for Systems Biology

Maintainer: <[e-mail]>

References

(1) Reiss, DJ and Facciotti, MT and Baliga, NS. (2007). "Model-based deconvolution of genome-wide DNA binding", Bioinformatics; doi: 10.1093/bioinformatics/btm592.
http://baliga.systemsbiology.net/medichi
(2) Qi, Y and et al. (2006). "High-resolution computational models of genome binding events", Nature Biotechnol, 24(8), 963-970. http://cgs.csail.mit.edu/jbd.

See Also

chip.deconv, deconv.entire.genome, fit.peak.profile, generate.fake.data, MeDiChI-data

Examples

## Compare profiles for DNA fragment distrs. with mean=300 and 400 bp.
kern.300 <- generate.binding.profile( fragment=function(x) dgamma( x,
                     shape=6, scale=50 ), verbose=TRUE )
kern.400 <- generate.binding.profile( fragment=function(x) dgamma( x,
                     shape=8, scale=50 ), verbose=TRUE )
plot( kern.300, typ="l", col="red" )
lines( kern.400, col="green" )

MeDiChI documentation built on May 2, 2019, 5:32 p.m.