ci: ci

Description Usage Arguments Details Value Author(s) See Also Examples

Description

Computes confidence intervals (CI) for Lambda using bootstrapping.

Usage

1
2
3
4
5
ci(object, subset, nBoot = 100, conf = 0.95, nCores = NULL, ...)

## S4 method for signature 'Les'
ci(object, subset, nBoot = 100, conf = 0.95, nCores =
NULL, ...)

Arguments

object

Object of class 'Les' as returned by 'estimate' or 'regions'.

subset

Vector of logical specifying the probes for which the CIs should be computed. If missing CIs will be computed for all probes.

nBoot

Integer specifying the number of bootstrap samples (default: 100). For details see 'boot' from the 'boot' package.

conf

Numeric specifying the confidence level (default: 0.95). For details see 'boot' from the 'boot' package.

nCores

Integer indicating the number of cores to use for computation. This feature requires the 'multicore' package which is only available for certain platforms. The package is used only if 'library(multicore)' has been called manually by the user before and if 'nCores' is an integer unequal NULL specifying the number of cores to use. The value is passed directly to 'mclapply' as argument 'n.cores'. For details and benefits please see the 'Details' section.

...

Further arguments passed to subsequent functions.

Details

The 'ci' method computes confidence intervals (CI) by bootstrapping probes in each window. Since based on percentiles the resulting CIs are asymmetrical.

All arguments for computation are taken from 'object' and thereby kept the same as for the results from 'estimation'.

Since bootstrapping is computational demanding and CIs are often only wanted for certain regions of interest it may be useful to restrict computation with the 'subset' argument.

The 'multicore' package can be used to spread the computation over several cores in a simple way. This can be useful on multi-core machines for large datasets or computation of confidence intervals for many probes. The 'multicore' package is not available on all platforms. To use multicore processing 'library(multicore)' has to be called beforehand and a number of cores to use has to be specified in 'nCores'. For details see the documentation of the 'multicore' package.

Value

Object of class 'Les' with additionally filled slots: ci, subset, nBoot, conf

Author(s)

Julian Gehring

Maintainer: Julian Gehring <julian.gehring@fdm.uni-freiburg.de>

See Also

Package: les-package boot

Class: Les

Methods and functions: Les estimate threshold regions ci chi2 export plot

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
data(spikeInStat)

x <- Les(pos, pval)
x <- estimate(x, win=200, grenander=FALSE)

subset <- pos >= 5232300 & pos <= 5233200
x <- ci(x, subset, conf=0.90, nBoot=50)

plot(x, error="ci")

## Not run: 
## multicore computation
## only available on certain platforms
library(multicore)
x <- ci(x, subset=150:200, conf=0.90, nBoot=50, nCores=2)

## End(Not run)

les documentation built on Nov. 8, 2020, 5:17 p.m.