Hi Christina. To be honest, I don't have much practical experience with the PSI, other than testing it for the book. I tried it on a few datasets and it seemed to produce reasonable-looking results. There are only a few papers out there using the PSI, one of which that comes to mind in this:
http://www.ncbi.nlm.nih.gov/pubmed/23365240
I'm very sorry the function wasn't in the zip file with the code. I don't know how that happened. I re-zipped the code with the data2psiX.m and re-posted it on my website. If you download the code again, you'll have the data2psiX.m function.