Résumé
Two amplified fragment length polymorphism (AFLP)-derived techniques were used to assess methylation at CCGG sites in the banana genome. Assessment of these techniques revealed that, while amplification steps are very reproducible, the ligation step is more subject to variability. Overall, these techniques produced an error rate of 0.2% per analysed band. Statistical approach highlights the fact that sample duplication is necessary to produce reliable results. This study involved 18 primer pairs and found that in the banana genome, roughly 80% of CCGG sites are unmethylated, 5% are methylated at the internal cytosine, and 15% are methylated at the external or both cytosines.