bltest: bltest

In lzcyzm/exomePeak: exome-based anlaysis of MeRIP-Seq data: peak calling and differential analysis

Description

This is the default test for the differential post-transcriptional RNA modification sites. Differential from all existing tests the compare the absolute amount between two conditions, this test compares whether the percentage of modified molecules are the same.

Usage

bltest(untreated_ip, untreated_input, 
       treated_ip, treated_input, 
       untreated_ip_total, untreated_input_total, 
       treated_ip_total, treated_input_total,
       minimal_count_fdr =10)

Arguments

untreated_ip	a vector of integers of n, which is the number of binding sites tested. Each element represents the number of reads fall into a binding site for the IP sample under untreated condition
untreated_input	a vector of integers of n, which is the number of binding sites tested. Each element represents the number of reads fall into a binding site for the Input control sample under untreated condition
treated_ip	a vector of integers of n, which is the number of binding sites tested. Each element represents the number of reads fall into a binding site for the IP sample under treated condition
treated_input	a vector of integers of n, which is the number of binding sites tested. Each element represents the number of reads fall into a binding site for the Input control sample under treated condition
untreated_ip_total	an integer, total number of reads for the IP sample under untreated condition
untreated_input_total	an integer, total number of reads for the Input control sample under untreated condition
treated_ip_total	an integer, total number of reads for the IP sample under treated condition
treated_input_total	an integer, total number of reads for the Input control sample under treated condition
minimal_count_fdr	an integer threshold, only the loci with reads more than this number are subjected for fdr calculation. default: 10

Details

The comparison of 4 Poisson distributions are firstly collapsed into 2 Binomial distributions, and the function further tests whether the two binomial distributions have the same successful rate with a likelihood ratio test. The number of reads at the same locus for the aligned reads are counted by other packages, such as Rsamtools or HTseq-count.

Value

The function returns a list of length 3, which contains the log(p-value), log(fdr) and log(fold change), respectively, from the test.

Author(s)

Lin Zhang, PhD <[e-mail]>

References

Reference coming soon!

Examples

# input reads count of 3 binding sites
untreated_ip = c(10,20,30)
untreated_input = c(20,20,20)
treated_ip = c(30,10,20)
treated_input  = c(20,20,20)
# sequencing depths
untreated_ip_total = 10^7
untreated_input_total = 10^7
treated_ip_total = 10^7
treated_input_total = 10^7
# get the result
result = bltest(untreated_ip, untreated_input, 
	treated_ip, treated_input, 
	untreated_ip_total, untreated_input_total, 
	treated_ip_total, treated_input_total)

lzcyzm/exomePeak documentation built on May 21, 2019, 9:16 a.m.