dodr: Detection of differences in rhythmic behavior between two...

In DODR: Detection of Differential Rhythmicity

Description

Detection of differences in rhythmic behavior between two time series sets

Usage

dodr(val1, val2, times1, times2 = times1, norm = TRUE, period = 24,
  method = "robust", verbose = options("verbose")[[1]])

Arguments

val1	matrix: data for values in first set. One column per sample, one row per time point
val2	matrix: data for values in second set. One column per sample, one row per time point
times1	vector: times of first set.
times2	vector: times of second set.
norm	boolean: whether to normalize the time series (division by mean), prior to the analysis.
period	numeric: period of the oscillations. Same unit as the time points in times1 and times2
method	vector<string>: method(s) to detect differences. Groups of related methods have additional identifiers. Elementary methods: HANOVA, harmScaleTest, harmNoisePred1, harmNoisePred2, robustDODR, robustHarmScaleTest) Groups: "all"All methods "robust"Combination of robust detection methods (robustDODR, robustHarmScaleTest) "lsq"Combination of least square based detetion methods (HANOVA, HarmScaleTest, HarmNoisePred1, HarmNoisePred2) "ANOVA"Combination of ANOVA like methods (HANOVA, robustDODR) "scaleTest"Combination of scaleTest methods (harmScaleTest, robustHarmScaleTest) "harmNoisePred"combination of both scaleTest methods to generate a two sided test
verbose	boolean: verbosity.

Details

This method applies a set of different methods on a pair of two experiments with one measurement matrix each. Samples to compare have to have the same column in both matrices. Different methods could be selected.

Value

A list containing

• 'p.value.table' A table containing the p-values for all the tests specified by method. Each row contains the results for one sample. A column meta.p.val is added containing the lowest p-value, corrected for multiple testing using a beta-distribution based aproach.

• details A list containing the detailed results from the selected methods

Examples

library(DODR)

#defining the parameters for two sets of oscillations
n=50
testTimes1 <- 0:15*3
testTimes2 <- testTimes1
tp <- length(testTimes1)
per1 <- 24
amp1 <- 0.3
ph1 <- 5
sd1 <- 0.1

per2 <- per1
amp2 <- amp1
ph2 <- ph1+4
sd2 <- sd1

#creating artificial oscillation sets
v1 <- 1 + amp1 * cos((testTimes1 - ph1)/per1*2*pi)
noise1 <- rnorm(length(testTimes1)*n, 0, sd1)
val1 <- matrix(v1 + noise1, ncol=n)

v2 <- 1 + amp2 * cos((testTimes2 - ph2)/per2*2*pi)
noise2 <- rnorm(length(testTimes2)*n, 0, sd2)
val2 <- matrix(v2 + noise2, ncol=n)

# run DODR
dodr <- dodr(val1, val2, testTimes1, testTimes2, 24, method = 'all')
dodr$p.value.table[1:3,]

#create another set with alterations in noise scale
ph2 <- ph1
sd2 <- sd1 * 3

v2 <- 1 + amp2 * cos((testTimes2 - ph2)/per2*2*pi)
noise2 <- rnorm(length(testTimes2)*n, 0, sd2)
val2 <- matrix(v2 + noise2, ncol=n)

dodr <- dodr(val1, val2, testTimes1, testTimes2, 24, method = 'all')
dodr$p.value.table[1:3,]

DODR documentation built on May 2, 2019, 8:33 a.m.