Differential Expression
- Once expression is normalised between samples, testing for between-group differences is in principle straightforward, and t-tests have been used for this purpose, but two factors complicate the analysis:
- Over-dispersion of variance
- usually accounted for using a negative binomial distribution
- Confounding factors
- usually accounted for with a linear modeling framework
- DESeq2: Likelihood ratio tests or Wald tests
- EdgeR: Liklihood ratio tests