A flow chart of the analytical process. Individual 'size class fingerprints' were firstly, (1) calibrated to size (bp) and then, (2) Fingerprints were assessed for gDNA contamination. (3) Data were assigned to a size continuum and (4) normalised so that heights of peaks between fingerprints could be compared. (5) The variability in each size class was measured amongst replicate PCRs, to establish significance levels. (6) Amplicon peaks were correlated with expected gene cassette amplicon sizes. (7) PCR biases were corrected to allow comparison of peaks within a fingerprint and the peaks representing gene cassettes rearranged into their order within the gene cassette array.