mbetattest: Performance of multiple beta t-test on simulated data
Description
This function is to peform multiple beta t-test method on real data. The result lists geneid or isoformid, gene name, the other information, t-value, p-value, adjusted p-value, adjusted alpha value, rho, and symb. All these lists are ordered by absolution of t-values.
Usage
mbetattest(X, na, nb, W, alpha, file)
Arguments
X
count data of RNA reads with na replicates in condition A ans nb replicates in condition B.
na
number of replicate libraries in condition A.
nb
number of replicate libraries in condition B.
W
numeric parameter, called omega that is a constant,determined by null simulation.
alpha
the probabilistic threshold. User can set alpha=0.05 or 0.01 or the other values. Defalt value is 0.05
file
a csv file. User needs to give file name and specify direction path. But if user uses setwd function, drive is not necessarily specified in file.
Value
return a dat list: the data ordered by abs(t) contain information cloumns, data columns, t-values, rho and symb that are used to make heatmap and MAplot.
Details
t-statistic is defined as t-statistic multiplied by (rho/omega), that is,
$$T=t*rho/omega$$
where
$$t=\frac{(PA-PB)}{sqrt(VA+VB)}$$
$$rho=sqrt(psi*zeta)$$
where
$$psi =max(\frac{min(XA)}{max(XB)},\frac{min(XB)}{max(XA)})$$
$$zeta=log(1+\frac{(mean(XA,XB)*var(XA,XB))}{(mean(XA)*var(XA)+mean(XB)*var(XB))})$$
omega is a constant as threshold.
References
Yuan-De Tan Anita M. Chandler, Arindam Chaudhury, and Joel R. Neilson(2015) A Powerful Statistical Approach for Large-scale Differential Transcription Analysis. Plos One, 10.1371/journal.pone.0123658.