I do not have access to suitable computational facilities to analyse raw reads of an RNAseq experiment.

Therefore I requested the raw read counts instead from the author of a paper in order to analyse the conditions myself. However the author has only provided me with the Deseq2 normalised counts. Is there a way for me to reverse calculate the raw read counts or proceed with analysis using Deseq2? If yes, which step in the vignette would I need to feed these counts into?

Ask them again for the original counts. It’s really just easier to get the correct data than to do computational backflips to avoid getting the correct input.