likelihood_ratio_poisson_exon_gene: Log Likelihood ratio statistic for comparison of the splicing...

In GPseq: gpseq: Using the generalized Poisson distribution to model sequence read counts from high throughput sequencing experiments

Description

Comparing the splicing rate of Exon (count vector : z in Tissue/Condition 1, v in Tissue/Condition 2) in a given Gene ( count vector : x in Tissue/Condition 1, y in Tissue/Condition 2). The counts are modeled as a Poisson Distribution with lambda as the Maximum Likelihood Estimate. The Log Likelihood ratio is distributed as a Chi-Square distribution with degrees of freedom 1. Newton Raphson method is used to estimate the parameters in the null model. The results are only applicable if the Netwon Raphson algorithm converges. The Null is that the exon has the same splicing rate in both conditions.

Usage

likelihood_ratio_poisson_exon_gene(z, x, v, y)

Arguments

z	Count vector of Exon in Tissue/Condition 1
x	Count vector of Gene in Tissue/Condition 1
v	Count vector of Exon in Tissue/Condition 2
y	Count vector of Gene in Tissue/Condition 2

Value

Ptest

The Chi square statistic = -2*Log Likelihood Ratio Statistic

Author(s)

Sudeep Srivastava, Liang Chen

References

Consul, P. C. (1989) Generalized Poisson Distributions: Properties and Applications. New York: Marcel Dekker.
Sudeep Srivastava, Liang Chen A two-parameter generalized Poisson model to improve the analysis of RNA-Seq data Nucleic Acids Research Advance Access published July 29,2010 doi : 10.1093/nar/gkq670

Examples

set.seed(666);
z = rpois(100,1);
x = c(z,rpois(200,5));
v = rpois(100,4);
y = c(v,rpois(200,8));

output = likelihood_ratio_poisson_exon_gene(z,x,v,y);

cat("Chi Square Statistic = ",output$Ptest,"\n");

GPseq documentation built on May 30, 2017, 3:11 a.m.