TRESS_DMRtest: Hypothesis testing on candidate DMRs.

In haowulab/TRESS: Toolbox for mRNA epigenetics sequencing analysis

Hypothesis testing on candidate DMRs.

Description

This function conducts statistical test for each candidate DMR based on user specified contrast of coefficients in design.

Usage

TRESS_DMRtest(DMR, contrast, nullModel = "standN")

Arguments

DMR	A list at least containing IP/input ratio, the coefficients estimate, variance-covariance estimate. This can be obtained from the ouput of TRESS_DMRfit.
contrast	A contrast for all coefficients in the design. It can be either a (p+1) vector or a m-by-(p+1) matrix, where p is the number of columns in the design. m depends on the number of relationships that users want to test.
nullModel	A character to specify a method to calculate p-value based on the statistics. It can be "standN", "2mix" and "trunN" for standard normal, two-component mixed gaussian and truncated normal respectively. Defult is "standN".

Details

The hypothesis for each of candidate DMR i is of the form:

H_0: C^TR_i = 0 vs. H_1: C^TR_i \neq 0

where C is a contrast of all coefficients in model design; R_i is coefficient vector for DMR i. If the C is a vector, then TRESS performs Wald test; if the C is a matrix, then TRESS conducts F-test.

Value

This function returns a dataframe containing the testing results for specified contrast. The columns are

baseMean	Averaged methylation level cross all samples.
logOR	Estimated value of contrast: C^TR_i. Only available if constrast is a vector.
lorSE	Standard error of C^TR_i. Only available if constrast is a vector.
stat	Test statistics.
pvalue	P-values from statistical tests.
padj	Benjamini-Hochberg procedure adjusted p-values.

Author(s)

Zhenxing Guo <[e-mail]>

References

Zhenxing Guo, Andrew M. Shafik, Peng Jin, Hao Wu. "Differential RNA Methylation Analysis for MeRIP-seq Data under General Experimental Design."

Examples

# A toy example
data(DMR_M3vsWT) # data from TRESS
variable = data.frame(predictor = rep(c("WT", "M3"), c(2, 2)))
model = ~1+predictor
DMR.fit = CallDMRs.paramEsti(
    counts = DMR_M3vsWT$Counts,
    sf = DMR_M3vsWT$sf,
    variable = variable,
    model = model
    )
DMR.test = TRESS_DMRtest(DMR = DMR.fit, contrast = c(0, 1))
head(DMR.test, 3)
head(DMR_M3vsWT$Regions[which(DMR.test$padj < 0.05), ], 3)

haowulab/TRESS documentation built on Aug. 27, 2022, 7:11 p.m.