bltest {exomePeak} | R Documentation |
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.
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)
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 |
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.
The function returns a list of length 3, which contains the log(p-value), log(fdr) and log(fold change), respectively, from the test.
Lin Zhang, PhD <laurenie.zhang@gmail.com>
Reference coming soon!
# 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)