Chicago (version 1.0.3)

getScores: Get CHiCAGO scores.

Description

Converts p-values into a CHiCAGO score, using p-value weighting.

Usage

getScores(cd, method = "weightedRelative",  includeTrans = TRUE, plot = TRUE, outfile = NULL)

Arguments

cd

A chicagoData object.

method

Either "weightedRelative" (recommended), or "unweighted".

includeTrans

If FALSE, trans interactions are discounted.

plot

Plot a diagnostic plot.

outfile

A string containing a .pdf file location to write to.

Value

An object of class chicagoData.

Warning

The object intData(cd) is updated by reference. Thus, intData(cd) will be altered. See vignette for further information.

Details

Weighting is performed using the parameters weightAlpha, weightBeta, weightGamma, weightDelta. Briefly, this function calculates weights w that decrease with increasing distance. Then, we construct weighted p-values p/w. As a result, the significance of long-range interactions is upweighted, and the significance of short-range interactions is downweighted.

Finally, the output score is calculated as -log(p/w) - log(w_max), where w_max is the highest attainable weight, and provided the score is positive (otherwise it is set to 0).

Please see the CHiCAGO paper and its supplementary for full details.

References

Genovese, C. R., Roeder, K., and Wasserman, L. (2006). False discovery control with p-value weighting. Biometrika, 93, 509-524. doi:10.1093/biomet/93.3.509

See Also

chicagoPipeline

Examples

data(cdUnitTest)

##modifications to cdUnitTest, ensuring it uses correct design directory
designDir <- file.path(system.file("extdata", package="Chicago"), "unitTestDesign")
cdUnitTest <- modifySettings(cd=cdUnitTest, designDir=designDir)

cdUnitTest <- getScores(cdUnitTest)