R: Compute moving average statistics by incorporating the...
cmarrt.ma
R Documentation
Compute moving average statistics by incorporating the correlation structure
Description
This function extends the moving average approach by incorporating the correlation structure. It also outputs the p-values of the standardized moving average statistics under the Gaussian approximation.
which chromosome should be analysed? If chr==NULL, all chromosome in the probeAnno object are taken.
M
rough estimate of the percentage of bound probes. If unknown, leave it NULL.
frag.length
average fragment length from sonication.
window.opt
option for sliding window, either "fixed.probe" or "fixed.gen.dist". Default is 'fixed.probe'.
Details
Computation using window.opt = "fixed.probe" calculates the moving average statistics within a fixed number of probes and is more efficient. Use this option if the tiling array is regular with approximately constant resolution. window.opt="fixed.gen.dist" computes the moving average statistics over a fixed genomic distance.
Value
data.sort
datafile sorted by genomic position.
ma
unstandardized moving average(MA) statistics.
z.cmarrt
standardized MA under correlation structure.
z.indep
standardized MA under independence (ignoring correlation structure).
pv.cmarrt
p-values of probes under correlation.
pv.indep
p-values of probes under independence (ignoring correlation structure).
Note
The p-values are obtained under the Gaussian approximation. Therefore, it is important to check the normal quantile-quantile plot if the Gaussian approximation is valid. The function also outputs the computation under independence (ignoring the correlation structure) for comparisons.
Author(s)
Pei Fen Kuan, Adam Hinz
References
P.F. Kuan, H. Chun, S. Keles (2008). CMARRT: A tool for the analysiz of ChIP-chip data from tiling arrays by incorporating the correlation structure. Pacific Symposium of Biocomputing13:515-526.