Résumé
A study was conducted to explore the genetic basis of amylose and amylopectin content in Dioscorea alata, key components influencing tuber quality. The goal was to identify quantitative trait loci (QTL) associated with these traits in two segregating populations as well as candidate genes. Phenotypic data were obtained using Near-Infrared Spectroscopy (NIRS). QTL analysis identified 14 QTLs, with 6 associated with amylopectin and 8 with amylose, distributed across 10 of the 20 yam chromosomes. Eight candidate genes involved in starch synthesis and/or degradation were identified, including glycosyltransferase, amylase synthase, and isoamylase ISA3. In parallel, 10 KASPar markers were developed from SNPs previously identified as being associated with several phenotypic traits influencing tuber quality, such as firmness, moldability, starch content, and cooking time. The results indicate that several of these markers could be used for marker-assisted selection (MAS), thereby facilitating the screening of traits that are difficult to phenotype in large segregating populations. Furthermore, efforts have been directed towards developing predictive models for pounded yam quality using NIRS, with a focus on moldability and amylose content. Additionally, work was carried out to standardize spectral databases across different generations of spectrometers, particularly by transferring the extensive dataset from the outdated FOSS NIRSystem 6500 to modern devices, including the Spectral Evolution NaturaSpec and ASD QualitySpec Trek.