Abstract
Accurate estimation of gene expression differences during development requires sensitive techniques combined with gold-standard normalization procedures. This is particularly true in the case of quantitative traits, where expression changes might be small. Nevertheless, systematic selection and validation of reference genes has been overlooked, even in Drosophila studies. Here, we tested the stability of six traditional reference genes across samples of imaginal wing disks from morphologically divergent strains of Drosophila melanogaster, in a two-class comparison: quantitative or qualitative variation in wing morphology. Overall, we identified and validated a pair of genes (RpL32 and Tbp) as being stably expressed in both experimental comparisons. These genes might be considered as a bona fide pair of reference genes for gene expression analyses of morphological divergence in D. melanogaster wings. They might also be taken as good candidates for experimental identification of stable reference genes in other morphological comparisons using Drosophila or other insect species. Besides, we found that some genes traditionally used as reference in qPCR experiments were not stably expressed in wing disks from the different fly strains. In fact, a significant bias was observed when the expression of three genes of interest, which are involved in the regulation of growth and patterning during imaginal wing development, was normalized with such putative reference genes. Our results demonstrate how inaccurate findings and opposite conclusions might be drawn if traditional reference genes are arbitrarily used for internal normalization without proper validation in the given experimental condition, a practice still common in qPCR experiments.
