Unraveling the Deep Genetic Architecture for Seedlessness in Grapevine and the Development and Validation of a New Set of Markers for VviAGL11-Based Gene-Assisted Selection.

Seedless inheritance has been considered a quasi-monogenic trait based on the VvAGL11 gene. An intragenic simple sequence repeat (SSR) marker, p3_VvAGL11, is currently used to opportunely discard seeded progeny, which represents up to 50% of seedlings to be established in the field. However, the rate of false positives remains significant, and this lack of accuracy might be due to a more complex genetic architecture, some intrinsic flaws of p3_VvAGL11, or potential recombination events between p3_VvAGL11 and the causal SNP located in the coding region. The purpose of this study was to update the genetic architecture of this trait in order to better understand its implications in breeding strategies. A total of 573 F1 individuals that segregate for seedlessness were genotyped with a 20K SNP chip and characterized phenotypically during four seasons for a fine QTL mapping analysis. Based on the molecular diversity of p3_VvAGL11 alleles, we redesigned this marker, and based on the causal SNP, we developed a qPCR-HRM marker for high-throughput and a Tetra-ARMS-PCR for simple predictive analyses. Up to 10 new QTLs were identified that describe the complex nature of seedlessness, corresponding to small but stable effects. The positive predictive value, based on VvAGL11 alone (0.647), was improved up to 0.814 when adding three small-effect QTLs in a multi-QTL additive model as a proof of concept. The new SSR, 5U_VviAGL11, is more informative and robust, and easier to analyze. However, we demonstrated that the association can be lost by intragenic recombination and that the e7_VviAGL11 SNP-based marker is thus more reliable and decreases the occurrence of false positives. This study highlights the bases of prediction failure based solely on a major gene and a reduced set of candidate genes, in addition to opportunities for molecular breeding following further and larger validation studies.