Using base R functions, you can generate a PCA bi- and tri-plot. See here: A: PCA plot from read count matrix from RNA-Seq
Using Z-scores is indeed most likely your best chance in terms of 'merging' these datasets and having them on the same distribution. However, in the case of RNA-seq, it's important to understand how you initially normalised the data. What I would do is normalise via EdgeR or DESeq2 and then transform the normalised counts via the rlog function of DESeq2, followed by a further transformation to the Z-scale. If you've got FPKM counts, then there is already a R function / package, called zFPKM, which claims to be able to transform FPKM to the Z scale too.
Edit (May 27, 2018): whilst it may be an interesting exercise to merge these types of data together, the statistical implications are not sound. The data are measured and processed differently and suffer different types of biases. We can get them onto th same data distribution (log2, Z-scale, etc), but the differences between the individual units on that distribution may not be comparable.