NBBttest (version 1.0.1)

betaparametVP: Estimation of parameters V and P in count data of RNA reads

Description

This function is used to estimate parameters P and V by optimalizing estimates of parameters: alpha and beta.

Usage

betaparametVP(X, NX)

Arguments

X

count dataset derived from m replicate libraries in one condition.

NX

vector of m library sizes. Library size is sum of counts over the whole library.

Value

return a list:

P

N proportions estimated.

V

N variances estimated.

Details

Count data of RNA sequence reads are assumed to follow binomial distribution with parameters (P) and (n) or negative binomial distribution with parameters (P) and (r) , while P is frequency of a gene or an isoform in RNA sequence population and assumed to follow beta distribution with parameters alpha (a) and beta(b). Parameters P and V are estimated by using optimal estimation of parameters a and b. The optimal method is an iteration algorithm drived by weighting proportion of gene or isoform in each replicate library. This is a large-scale method for estimating these parameters. Estimation of parameters P and V is core of the multiple beta t-test method because P and V will be used to calculate t-value.

References

Baggerly KA, Deng L, Morris JS, Aldaz CM (2003) Differential expression in SAGE: accounting for normal between-library variation. Bioinformatics, 19: 1477-1483.

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.

See Also

betaparametab, betaparametw

Examples

# NOT RUN {
data(jkttcell) 
X<-jkttcell[1:100,]
na<-3
nb<-3
cn<-length(X[1,])
rn<-length(X[,1])
XC<-X[,1:(cn-na-nb)]
XX<-X[,(cn-na-nb+1):cn]
n<-na+nb
XA<-XX[,1:na]
SA<-apply(XA,2,sum)
PA<-betaparametVP(XA,SA)
# }