ctest {exomePeak2} | R Documentation |
Exact poisson test on the ratio between IP and input counts.
ctest( IP_count, input_count, IP_sizeFactor, input_sizeFactor, fold = 1, alpha = 0.05 )
IP_count |
a numeric vector for the counts of IP sample. |
input_count |
a numeric vector for the counts of input sample. |
IP_sizeFactor |
a numeric vector for the size factors of IP sample. |
input_sizeFactor |
a numeric vector for the size factors of input sample. |
fold |
the fold change under the null hypothesis; default = 1. |
alpha |
the alpha level used for determining the optimal independent filtering in |
C-tests will be conducted between each entries of the IP and input vector.
The following statistical model is assumed on the data:
IP ~ Poisson(mean_IP*IP_sizeFactor)
input ~ Poisson(mean_input*input_sizeFactor)
The one-sided statistical test (C-test) is conducted with the following hypothesis pair:
H_0: mean(IP)/mean(input) <= fold
H_1: mean(IP)/mean(input) > fold
The exact p-values will be generated using binomial test, check poisson.test
for more details.
The p-values are adjusted with "BH" method with an independent filtering slected by function filtered_R
in package genefilter
;
the filter statistics used is the normalized read abundence.
A list include adjusted p values (with method fdr) and the corresponding log2 fold changes.