Likelihood
Ratio Test Statistic (E2)
There
are three methods for computing E2 in the ISOGeneGUI. First is
asymptotic, second is using permutations and last is using
faster implementation of permutation testing from package
orQA.
1.
Asymptotic method
This options provides the analysis using the E2 test statistics
based on p-values obtained from its exact distribution. The provided
p-values adjustment methods are BH, BY, Holm, SidakSS, Sidak SD,
Holm, Hochberg and Bonferroni. Based on the specified adjustment
methods and overall significant level, the list of significant genes
along isotonic means difference, test statistic value, raw and
adjusted p-values, will be produced.
Analysis > Likelihood Ratio Test (E2): Asymptotic
Screenshot
of the main dialog box of Likelihood Ratio Test (E2) analysis:
Asymptotic
Screenshot
of the list of significant genes
Plots
After the raw p-values and adjusted p-values are calculated, there
are three plot that can be produced. After choose the option and
click the Produce
plots button
the plot will be displayed.
Screenshot
of the plot of Fold change vs. E2
3.
Permutation (orQA)
Last
option is using implementation from package orQA. The function
e2test is used. The function is originally aimed for resampling
within one mouse (in the context of experiment described in orQA
package). However, we are using the method in the way that
resampling is done across all observation, i.e. considering all our
observations independent.
Screenshot
of the main dialog box of Likelihood Ratio Test (E2) analysis: orQA
Plots
After the raw p-values and adjusted p-values are calculated, there
are two plot that can be produced. After choose the option and click
the Produce plots button the plot will be displayed.
Screenshot
of the plot of the p-values
