LR_rhythmicity: Likelihood-based Tests for Detecting Circadian Pattern.

In diffCircadian/diffCircadian: An R package for circadian and differential circadian analysis.

Likelihood-based Tests for Detecting Circadian Pattern.

Description

Likelihood-based tests for circadian pattern detection.

Usage

LR_rhythmicity(tt, yy, period = 24, method = "LR", FN = TRUE)

Arguments

tt	Time vector
yy	Expression vector
period	Period of the since curve. Default is 24.
method	Testing methods can be "Wald" or "LR". Default is "LR".
FN	Type of Test, finite sample if TRUE or large sample if FALSE. Default is TRUE.

Details

Test the significance of circadian curve fitting using likelihood-based tests.

Value

A list of amp, phase, offset, sigma02, sigmaA2, l0, l1, df, stat, and pvalue. Formula 1: yy = amp * sin(2π/period * (phase + tt)) + offset Formula 2: yy = A * sin(2π/period * tt) + B * cos(2*π/period * tt) + offset

amp	Amplitude based on formula 1.
phase	Phase based on formula 1, phase is restricted within (0, period).
peakTime	Phase based on formula 1, peakTime is restricted within (0, period). phase + peakTime = period/4
offset	Basal level (vertical shift) based on formula 1 or on formula 2.
sigma02	Variance estimate under the null (intercept only).
sigmaA2	Variance estimate under the alternative (since curve fitting).
l0	Log likelihood under the null (intercept only).
l1	Log likelihood under the alternative (since curve fitting).
stat	Test statistic.
pvalue	P-value from the test.
R2	Pseudo R2 defined as (tss - rss)/tss.

Author(s)

Zhiguang Huo, Haocheng Ding

Examples

set.seed(32608)
n <- 50
tt <- runif(n,0,24) 
Amp <- 2
Phase <- 6
Offset <- 3
yy <- Amp * sin(2*pi/24 * (tt + Phase)) + Offset + rnorm(n,0,1)
LR_rhythmicity(tt, yy, period=24, method="LR", FN=TRUE)

diffCircadian/diffCircadian documentation built on Feb. 24, 2023, 9:07 a.m.