estimateThreshold: EstimateWindowThresholds

In christpa/DMRScan: Detection of Differentially Methylated Regions

EstimateWindowThresholds

Description

Estimate window thresholds for sliding window, one unique value for each window size

Usage

estimateWindowThreshold(nProbe, windowSize, method = "siegmund",
  mcmc = 1000, nCPU = 1, submethod = "ar", ...)

Arguments

nProbe	The number of probes (CpGs) in the study.
windowSize	The different window sizes to be tested. Must be either one, or an ordered sequence of integers.
method	Gives the method by which the threshold is calculated. Can be either an analytical solution "siegmund", provided by Siegnumd et.al (2012), or an iterative process; either importance sampling "sampling", as suggested by Zhang (2012) or a full MCMC model "mcmc" which can account for any dependency structure, wich is pass to arima.sim, with ...
mcmc	The number of MCMC iterations to be used, when using either Important Sampling ("zhang") or MCMC estimation of the threshold.
nCPU	When calculating the thresholds on a cluster, how many CPUs should be used. This option is only compatible with the 'mcmc' method.
submethod	A character string indicating if an AR(5) or ARIMA model should be used. In the AR(5), the index runs from -2 to 2. A regular AR(p) model can be obtaine using ARIMA(p,0,0) instead.
...	Optinal parameters pased on to arima, when simulating data using the mcmc option, see arima.sim()

Value

Returns a vector of the threshold for each window size

Examples

thresholdGrid <- estimateWindowThreshold(nProbe = 1000, 
                                    windowSize = 3:8, method = "siegmund")

christpa/DMRScan documentation built on Nov. 16, 2024, 9:36 a.m.