LRtest1: Likelihood ratio test for differential expression analysis...

In zhanghfd/PLNseq: PLNseq: A multivariate Poisson lognormal distribution for high-throughput correlated RNA-sequencing read count data

Description

This function calculates log-fold changes, likelihood ratio test statistics, and p-values for a list of genes. This function should be called after a commom correlation matrix is returned by ‘correlationCoefficient’.

Usage

    LRtest1(d,z,use.commonSigma,id)

Arguments

d	This is a PLNseq object.
z	J independent samples (a matrix of dimension J by R) drown from multivariate normal distribution with expectations 0, variances 1, and a common correlation coefficient matrix estimated by ‘correlationCoefficient’.
use.commonSigma	Use common ‘sigma’ (TRUE) or genewise ‘sigma’ (FALSE), with default value ‘FALSE’.
id	A vector consisting of a subset of 1,...,J, with default value 1:J (all genes are analyzed).

Value

LR	Estimation and test results: ‘log-FC’, ‘LR statistic’, ‘p value’.

Examples

    data(lung);
    count = lung[,c(2:4,8:10)];
    d = PLNobject(count,conditionNumber=2);
    d = sizeFactor(d,maxCount=2e3);

    ##  Not run:
    ##  d = commonSigma(d);
    ##  d$commonCorrelation = TRUE;
    ##  d = correlationCoefficient(d);
    ##  d = genewiseSigma(d);
    ##  library('MASS');
    ##  z = mvrnorm(n=1e5,mu=rep(0,2),Sigma=d$rho);
    ##  d = LRtest1(d,z,use.commonSigma=FALSE,id=1:100);

zhanghfd/PLNseq documentation built on May 4, 2019, 10:16 p.m.