Key message This study generated the first high-density genetic map for D. alata based on genotyping-by-sequencing and provides new insight on sex determination in yam. Abstract Greater yam ( Dioscorea alata L.) is a major staple food in tropical and subtropical areas. This study aimed to produce the first reference genetic map of this dioecious species using genotyping-by-sequencing. In this high-density map combining information of two F1 outcrossed populations, 20 linkage groups were resolved as expected and 1579 polymorphic markers were ordered. The consensus map length was 2613.5 cM with an average SNP interval of 1.68 cM. An XX/XY sex determination system was identified on LG6 via the study of sex ratio, homology of parental linkage groups and the identification of a major QTL for sex determination. Homology with the sequenced D. rotundata is described, and the median physical distance between SNPs was estimated at 139.1 kb. The effects of segregation distortion and the presence of heteromorphic sex chromosomes are discussed. This D. alata linkage map associated with the available genomic resources will facilitate quantitative trait mapping, marker-assisted selection and evolutionary studies in the important yet scarcely studied yam species. Electronic supplementary material The online version of this article (10.1007/s00122-019-03311-6) contains supplementary material, which is available to authorized users.

Dioscorea alata is a polyploid species with several ploidy levels and its basic chromosome number has been considered by most authors to be x = 10. Standard chromosome counting and flow cytometry analysis were used to determine the chromosome number of 110 D. alata accessions of the CIRAD germplasm collection. The results revealed that 76% of accessions have 2n = 40 chromosomes, 7% have 2n = 60 chromosomes and 17% have 2n = 80 chromosomes. Progenies were produced from 2n = 40 types of D. alata and the segregation patterns of six microsatellite markers in four different progenies were analysed. The Bayesian method was used to test for diploid versus tetraploid (allo- and autotetraploid) modes of inheritance. The results provided the genetic evidence to establish the diploidy of plants with 2n = 40 chromosomes and to support the hypothesis that plants with 2n = 40, 60 and 80 chromosomes are diploids, triploids and tetraploids, respectively, and that the basic chromosome number of D. alata is x = 20. The findings obtained in the present study are significant for effective breeding programs, genetic diversity analysis and elucidation of the phylogeny and the species origin of D. alata.

Two Dioscorea alata populations were generated by hand pollination between contrasted diploid genitors. Population A (74F × Kabusa) was composed of 121 progenies while population B (74F × 14M) involved 193 progenies. These two populations were assessed over two consecutive years regarding important tuber quality traits. Analysis of variance showed that the genotype had the greatest influence on the phenotypic scores. Also for some traits, effect of the year_replicate was strong. The heritabilities of most traits were high. Based on these data and a reference high-density genetic map of greater yam, a total of 34 quantitative trait loci (QTLs) were detected on 8 of the 20 yam chromosomes. They corresponded to five of each of the following traits: tuber size, shape regularity, tubercular roots, skin texture, tuber flesh oxidation, six for oxidation ratio and three for flesh colour. The fraction of total phenotypic variance attributable to a single QTL ranged from 11.1 to 43.5%. We detected significant correlations between traits and QTL colocalizations that were consistent with these correlations. A majority of QTLs (62%) were found on linkage group LG16, indicating that this chromosome could play a major role in genetic control of the investigated traits. In addition, an inversion involving this chromosome was detected in the Kabusa male. Nine QTLs were validated on a diversity panel, including three for tuber size, three for shape regularity, two for skin texture and one for tubercular roots. The approximate physical localization of validated QTLs allowed the identification of various candidates genes. The validated QTLs should be useful for breeding programs using marker-assisted selection to improve yam tuber quality.