A simple AFLP model, using the relatively small bacteriophage lambda genome, was developed to test the reproducibility of this technique in an international study, CCQM-P53. Using either non-selective or selective primers, 9 fragments or a subset of 1 to 3 fragments, respectively, were predicted using in silico software. Under optimized conditions, all predicted fragments were experimentally generated. The reproducibility of the AFLP model was tested by submitting both unknown DNA template which had been restricted and ligated with AFLP linkers (R/L mixture) and corresponding primer pairs to 9 laboratories participating in the CCQM-P53 study. Participants completed the final PCR step and then used either slab gel electrophoresis or CE to detect the AFLP fragments. The predicted fragments were identified by the majority of participants with size estimates consistently up to 4 base pair (bp) larger for slab gel electrophoresis than for CE. Shadow fragments which were 3 bp larger than predicted fragments were often observed by both study participants and organizers. The 9 AFLP fragments exhibited consistent differences in peak height and reproducibility in the CE profiles with fragments containing the highest guanine-cytosine (GC) content of 50-56% showing the greatest stability in the AFLP profiles.
Pub Type: Journals
AFLP, amplification efficiency, comparability, GC content, lambda phange, reproducibility