I have a set of 96 single-cell RNA-Seq samples. They are all of the same type of cells but I would like to see whether there is any hidden structure in this cell population. I was thinking of performing PCA. I wanted to use DESeq2 but their workflow seems to involve computing the differential expression first:

ddsHTSeq<-DESeqDataSetFromHTSeqCount(sampleTable=sampleTable, directory=getwd(), design=~condition)

The problem is that I don't have a "condition" field. All samples are from the same population. It is not known whether there is any difference between them. So, I would like to compute a PCA which does not involve or depend on computing differential expression between 2 or more groups. How can I do this? If DESeq2 can't do this can you recommend an alternative?

You can use ~1.

This is what the transformations do internally as well by default, see the 'blind' argument.