For downstream fragment analysis (T-RFLP), I am running PCRs with primers labelled with the fluorescent tags HEX and 6-FAM. After having some trouble gaining enough product, I now do a primary PCR with the same primers but unlabelled, then a secondary PCR with the labelled primers. The amount of product may be sufficient, but possibly not. Do you have any advice or experience regarding the optimisation of such a PCR protocol?