Dear Rasmus,
We are working on exactly this problem (analyze Mo lab-data using pdfgetX3). Please can you reach out directly to me (
sb2...@columbia.edu) and I can put you in touch with the team of students working on this..
For everyone in general, PDFgetX3 will process lab data to get a nice PDF, no problem, but a key assumption behind the algorithm is violated in many cases, that of negligible absorption, so although the PDF you get will look good, it will be wrong. We are working on a fix for this. Please stay tuned.