Table 5.

Methods for identifying differentially expressed genes in individual samples.

Software	Description	Data Type	Year	Ref.
GFOLD	Combines fold change and statistical significance, assumes Poisson distribution in the absence of biological replicates and estimates uncertainty.	Paired-Tumour/Normal	2012	Feng et al. [65]
DESeq2	In the absence of biological replicates, assumes that most genes will not be differentially expressed, and uses the two conditions (tumour/normal) as their own replicates to calculate mean variance.	Paired-Tumour/Normal	2014	Love et al. [47]
RankComp	Creates ranked-ordered list of genes, looks for gene pairs with stable ordering across reference samples and then finds genes with reversed order.	Tumour Only	2015	Wang et al. [66]
NOISeq-Sim	Simulates technical replicates assuming a multinomial distribution. Only a simulation of technical replicates.	Paired-Tumour/Normal	2015	Tarazona et al. [64]
PenDA	Creates ranked-ordered list of genes, compares local ordering of a gene in a sample of interest versus a set of reference samples.	Tumour Only	2020	Richard et al. [67]