Table 2. Description of the core modeling strategy of differential gene expression analysis methods investigated in the present study.

Method	Description	Reference
NOISeq	Non-parametric modeling of odds of signal against noise;	Tarazona et al. (2011)
	NOISeqBIO is a variant for handling replicated experiments which	Tarazona et al. (2015)
	integrates the non-parametric framework of NOISeq with an
	empirical Bayes approach
ASC	Empirical Bayes shrinkage estimation of log fold change	Wu et al. (2010)
GFOLD	Poisson count distribution; Bayesian posterior distribution	Feng et al. (2012)
	for log fold change
edgeR	Negative binomial count distribution; genewise dispersion parameter estimation via conditional maximum likelihood; empirical Bayes shrinkage of dispersion parameter; exact test for p-value computation	Robinson, McCarthy & Smyth (2010)
DESeq	Negative binomial count distribution; local regression modeling of mean and variance parameters	Anders & Huber (2010)
DESeq2	Negative binomial count distribution; generalized linear model; shrinkage estimation of dispersion parameter and fold change	Love, Huber & Anders (2014)
voom	Estimates of mean–variance trend from log-transformed	Law et al. (2014)
	count data are used as input for the limma empirical Bayes
	analysis pipeline developed for microarray data analysis
Z-test	The Z-statistic for testing the equality of two proportions	–