Differential expression when working without any replicates

[bel012]

Hi,

This is a question regarding the DESeq (ver 1 or 2) package.
I wanted to know if we need to call estimateSizeFactors() when no replicates are available.
In the DESeq documentation (ver 1 ) , in the section of
"Working without any replicates", the estimateSizeFactors() is skipped .
I could also see a post which suggested to work with rlogTransformation() to estimate the rLogFC (no pvalues), where no count normalization was done.
In the estimateSizeFactors documentation - it is written that the normalization is done over all samples (and not conditions).

Can anyone elaborate?

Thanks
bel.