Abstract
Introduction: Useful germplasm for citrus breeding includes all sexually compatible species of the former genera Citrus, Clymenia, Eremocitrus, Fortunella, Microcitrus, Oxanthera, and Poncirus, now merged in the single Citrus genus. An improved knowledge on the synteny/collinearity between the genome of these different species, and on their recombination landscapes, is essential to optimize interspecific breeding schemes. Method: We have performed a large comparative genetic mapping study including several main clades of the Citrus genus. It concerns five species (C. maxima, C. medica, C. reticulata, C. trifoliata and C. glauca), two horticultural groups resulting from interspecific admixture (clementine and lemon) and two recent interspecific hybrids (C. australis x C. australasica and C. maxima x C. reticulata). The nine individual genetic maps were established from GBS data of 1,216 hybrids. Results and discussion: The number of SNPs mapped for each parent varies from 760 for C. medica to 4,436 for the C. maxima x C. reticulata hybrid, with an average of 2,162.3 markers by map. Their comparison with C. clementina v1.0 assembly and inter-map comparisons revealed a high synteny and collinearity between the nine genetic maps. Non-Mendelian segregation was frequent and specific for each parental combination. The recombination landscape was similar for the nine mapped parents, and large genomic regions with very low recombination were identified. A consensus genetic map was successfully established. It encompasses 10,756 loci, including 7,915 gene-based markers and 2,841 non-genic SNPs. The anchoring of the consensus map on 15 published citrus chromosome-scale genome assemblies revealed a high synteny and collinearity for the most recent assemblies, whereas discrepancies were observed for some older ones. Large structural variations do not seem to have played a major role in the differentiation of the main species of the Citrus genus. The consensus genetic map is a useful tool to check the accuracy of genome assemblies, identify large structural variation and focus on analyzing potential relationships with phenotypic variations. It should also be a reference framework to integrate the positions of QTLs and useful genes identified in different analyses.