Résumé
SSRs have been considered as almost ideal markers for genetic diversity analysis. With the increasing availability of sequencing data, SNPs and InDels become major classes of codominant markers with genome wide coverage. We have analyzed the respective values of SSRs, InDels, and SNPs for intra and interspecific Citrus genetic diversity analysis. Moreover, we have compared the diversity structure revealed by markers mined in a single heterozygous genotype (clementine) and markers mined from a large interspecific survey. A random set of markers was selected for each marker class to genotype 48 citrus accessions. SSRs were the most polymorphic markers at the intraspecific level allowing complete varietal differentiation within basic taxa (C. reticulata, C. maxima, C. medica). However, SSRs gave the lowest values for interspecific differentiation, followed by SNPs and InDels, that in average displayed low intraspecific variability but high interspecific differentiation. A clear effect of the discovery panel was observed for SNPs and InDels. The ascertainment biases associated with the clementine heterozygosity mining resulted mainly in an over estimation of within C. reticulata diversity and an underestimation of the interspecific differentiation. In conclusion pluri-allelic SSRs are very useful for intraspecific structure analysis but have limitations for interspecific and phylogenetic studies. SNPs and InDels mined in large discovery panels appear to allow efficient selection of more specialized markers with high potential for phylogenetic diagnosis and to decipher the interspecific mosaic structure of secondary cultivated species or to analyze intraspecific diversity. Several SNP genotyping methods are available for different analysis scales, from high throughput genotyping to small set of data production. With the ongoing resequencing projects, SNPs markers will soon be the most important class of markers for citrus genetics.