An odd situation: rfu values & scoring of AFLP fragments in clonally propagated perennial crop
I have detected significant AFLP variability among samples where it should not be expected: clonally propagated young trees. How to set rfu values in scoring AFLP fragments?
please, help me to explain an odd situation regarding an AFLP analisis of perennial forestry crop and reading of fragments in AppliedBiosystems capillary ABI 3130. The situation is following:
- several samples of DNA from the clonal progenies of the same tree, all the trees are young, grown at the same location
- AFLP profiles, as expecting, should be the same: no mutations in short time in the young trees after vegetative propagation
- BUT, they are not the same: there are significant differences detected by AFLP.
I guess, there are some catch in the reading of aflp profiles: rfu values. It is common to read just fragments that are stronger than 50 rfu. However, I have a situation when, at the same position, I have peaks significantly stronger than 50 rfu, but also (at the another sample that should be the genetically the same) I have fragments of 10, 20 or 30 rfu, but also those of 300-500-900 rfu. The average strength of peaks might be 500 rfu, but there are also peaks of 40, 50 or 60 and 70 rfu. Should I eliminate all weak fragments, those of 60-70 rfu, regardless the limit value od 50 rfu?
If I eliminate these ''weak'' fragments and leave strong (more than 50 rfu), then I am in the odd situation, generating ''artificial genetic variability'' among the samples where any variability cannot be expected, and where no variability might be detectable by AFLP. At least by this range and in this frequency. How could you explain this?
The same happen if I set a limit, for example, at 100 or 150 rfu.
The key question is: how to determine a lower limit of rfu values?
What is your opinion about ScanAFLP software for fragment scoring in the case of vegetativelly propagated plants? Is ScanAFLP good enough for analyses of 20 - 30 clonal samples, just as good as if applied for large population analyses of cross-pollinated plants? Are there any alternatives?
I just add: I am sure there are no mistakes in sample mislabeling, no mistakes in pipetting, and I included repetition and blank samples i.e. pcr with pure water, without DNA.
I really appreciate your contribution in solving this mistery!