Hi,

We are relatively new to RNA-seq data analysis and we have some questions regarding DESeq2 although we have extensively read the paper and the vignette.

1)we wonder how and when DESeq2 takes specified covariates into account during the analysis. We also have to input batch effect and different groups to compare, so we would like to validate if the step that takes covariates into account is the Gene-wise dispersion estimate.

2)We would also like to know where the LFC comes from, especially since we have a continuous variable in input and you can’t calculate an «actual» fold change from this type of variable. Also we would like to know if there is a function in DESeq2 where we can get «intermediate» output file from different analysis steps (other than the mcols function) like the first LFC that we get from DESeq2 before the shrinkage.

3)We also have a question about how the independent filtering works (i.e. what are the principles behind it)?

Thank you very much for your help

DESeq2 uses all the specified covariates when estimating the dispersion (all steps), and obviously when estimating the coefficients (in the paper, the beta's, also described as log fold changes).

The LFC is the beta associated with that covariate. It is estimated the same way regardless of whether the covariate is continuous or discrete.

Take a look at the section of the vignette "Access to all calculated values"

Take a look also at the independent filtering section of the vignette, and the reference there

https://bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#independent-filtering-and-multiple-testing