Interactive demo · Statistical Methods for Modern Biology · Chapter 8

Dispersion shrinkage and variance stabilisation

With seven samples a gene-wise dispersion estimate rests on five degrees of freedom and is very noisy. DESeq2 maximises the Cox-Reid adjusted likelihood for each gene, fits a trend α(μ) = a0 + a1/μ across genes, and takes the posterior mode under a log-normal prior centred on the trend. The same trend gives a closed-form variance-stabilising transformation whose standard deviation is about √a0/ln 2 at every mean, unlike log2(n + n0).

View

Prior variance σd2 of log dispersions

Pseudocount n0 in log2(n + n0)

Simulated genes per mean

Readout

Try these

In the lecture notes