Sharka is one of the most serious viral diseases affecting stone fruit species and, in apricot, resistance to its viral agent, the Plum Pox Virus (PPV), is conferred by one major quantitative trait locus (QTL), named PPVres for PPV resistance. Previous studies indicated that PPV-resistant cultivars and breeding progenies can be selected by using a set of SSR markers (named PGS) targeting the PPVres locus. However, before these markers can be employed for marker-assisted selection, they were validated in a wide range of genetic backgrounds and environments. We used a total of 11 mapping populations issued from three distinct environments to confirm that this marker set located within the QTL adequately predicted PPV resistance. In this study, we show that selection of PPV-resistant material based only on markers co-localizing with the PPVres major locus is not fully reliable. Indeed, genotype-phenotype discrepancies were observed depending on the progeny and the PPV-resistant/susceptible parents. While most of the PPV-resistant individuals displayed the resistant alleles, a significant number of PPV-susceptible individuals showed the same resistant haplotype. An effect of the PPV strain used for phenotyping was also demonstrated. We thus hypothesize that the presence of other factors or genes involved in the mechanism of resistance to sharka in apricot could explain these unexpected results. Our work indicates that the current PGS marker set is not broadly applicable for MAS and that marker-assisted breeding based on the sole PPVres locus is not sufficient to unambiguously select PPV-resistant apricot cultivars.< Réduire