Cultivated grapevine (Vitis vinifera) is one of the most widely grown fruit crops in the world and held in high regard for its nourishing fruits, sweet juices and iconic wines. Global viticulture predominantly utilizes Vitis vinifera varieties, because they convey sensory attributes corresponding to the current consumer ideal of product quality. However, they are also highly susceptible to fungal pathogens, and therefore require intense applications of plant protection products with adverse side effects. Consumers criticize the use of pesticides for food production but simultaneously request perfect product quality. Viticulture could prove it is possible to reduce the demand for pesticides while keeping high quality standards by introducing newly bred varieties with resistances against downy and powdery mildew, two main fungal threats. Nevertheless, plant–pathogen interactions are cycles of resistance and susceptibility, and some strains of these pathogens have developed mechanisms to overcome the resistance within a few decades. Recently, grapevine breeding has started drawing on trait-linked molecular markers to combine several resistance loci within new cultivars for more endurable resistance. For grey mold, a third severe threat in viticulture, an active resistance mechanism is still not feasible. Therefore, grapevine-breeding aims at introducing fungi-static physical properties, e.g. wax layers, more or rigid cells in the berry skin and loose cluster architecture as additional defense mechanisms. This is a way to reduce the susceptibility to pathogens in general and in particular if physical resilience is the only effective option. The central hub of these physical barriers is a loosely clustered variety. The enhanced available space between the berries provides the framework for the effective formation of a firm berry surface and waxy cover and is restricting the time-span with favorable moisture conditions for fungal infections even inside of the cluster. The overall aim of this thesis was to shed light on genetic cues involved in cluster architecture and to derive first molecular markers that have the capacity to differentiate between loosely and compactly clustered genotypes. This provides the prerequisite for MAS of the desired loosely clustered individuals. To this end, the experimental design of this thesis draws on different sources of natural variance: Firstly, the F1 generation of the cross (‘Calardis Musqué’ × ‘Villard Blanc’) and secondly, somatic variants of the variety ‘Pinot Noir’ showing significantly different cluster compactness. Both sources of natural variation were successfully used to elucidate cluster architecture sub-traits that trigger phenotypic differences between loose and compact clusters. The genetic approach, applied in Chapter 2, exposed overlapping regions with up to four QTLs for cluster architecture sub-traits that are physically co-located on the grapevine reference genome. Based on co-location on the chromosome, this finding provides the option for a joint introgression of multiple genetic variations in a breeding scheme with an overall considerable effect on CA. In addition, several molecular markers with strong linkage to these cluster architecture sub-traits could be proposed (Richter et al. 2019). A ‘proof of concept’ study (Chapter 4) showed that it was possible to exploit three of these markers for MAS against unwanted compactly clustered individuals. This demonstrates their capacity as selective markers for a complex morphological trait among the individuals of the cross (‘Calardis Musqué’ × ‘Villard Blanc’). The survey in Chapter 3 reveals that the gene expression of 15 candidate genes consistently correlates to cluster architecture variations of ‘Pinot Noir’ clones in a multi environmental experiment. The genetic approach applied in Richter et al. (2019), the gene expression experiments in Richter et al. (2020) and the results of the RNA-sequencing previously described in Rossmann et al. (2020) provide multiple lines of evidence for the reported candidate genes. In further phenotypically divergent individuals from a genetically diverse background, the transcription factor gene PRE6 and six genes related to auxin metabolism, cell wall loosening and strigolactones showed differential expression (Richter et al. 2020). Implementing an evidence-based network, allowing a wider view on the interaction of the candidate genes, shows multiple associations of the candidate genes with brassinosteroids, a class of growth-promoting phytohormones. Thus, the candidate genes presented here may have the capacity to be successfully involved in marker development with the aim of selecting cluster architecture traits in MAS enabling breeders to identify optimized breeding material with physical resilience to fungal pathogens such as B. cinereaweiterlesen