Submitted to: Journal of Cereal Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/8/2017
Publication Date: 6/19/2017
Citation: Kiszonas, A., Boehm, J.D., See, D.R., Morris, C.F. 2017. Identification of SNPs, QTLs, and dominant markers associated with wheat flavor preference using genotyping-by-sequencing. Journal of Cereal Science. 76:140-147.
Interpretive Summary: Objectionable flavor and sensory perception of whole wheat products are often invoked as constraints to increased consumption, despite the documented health benefits associated with the consumption of whole wheat foods. Including the bran in whole wheat flour (in contrast to removing it in refined, white flour) is generally associated with stronger flavors which many consumers may find objectionable. The genetic basis for flavor differences in wheat are largely unknown. A major obstacle to understanding the underlying genetics, such that wheat cultivars with more desirable flavor profiles might be developed, is the difficulty in “phenotyping” germplasm. Most directly, this activity would be performed by trained human sensory panels under the careful guidance of sensory scientists. This approach, although ‘best’ from a theoretical standpoint, is expensive, labor and time intensive, limited in the number and scope of samples, and presents some ethical considerations. Our group has developed an alternative involving the laboratory mouse (Mus musculus L.) as a model system. The key features of this system include: the use of females of the C57BL/6 inbred mouse strain that are highly reproducible, replicate mice that provide an accurate estimate of error variance, binary mixtures of two wheat varieties or grain lots, whole wheat kernels without the confounding effects of further processing, and perhaps most importantly, the generation and application of Student’s t statistic as a consumption preference phenotype. Student’s t statistic is unitless, continuously distributed, and sign independent; it accounts for variance and sample size (replication). When used in conjunction with a common check variety, it provides a quantitative phenotype for variety ranking, mean separation, and genetic mapping. Using this mouse model system, our consumption preference results were found to be consistent with prior human studies: in general red bran wheats are consumed less than white bran wheats, however there were large differences among varieties independent of bran color. In our most recent report, we used this model system to identify quantitative trait loci (QTL) present in a hard white by soft white recombinant inbred line (RIL) population with a limited number of RILs (78) and limited marker coverage. No knowledge of the relative consumption preference of the parents was known a priori. After accounting for phenotypic and genetic grain hardness in the analysis, 32 significant QTL were identified.
Technical Abstract: Whole wheat products are well known to provide important nutrients in the human diet; however, consumer acceptance can be hindered by the flavor, aroma, and texture of whole wheat products. Flavor differences among wheat varieties have been observed, but are still little understood. A laboratory mouse model was used to examine flavor using a two-choice feeding system and the Student’s t statistic. Whole grain was used to eliminate the confounding effect of processing. The Student’s t statistic has been used previously to identify “Yummy” and “yucky” varieties. This study used next-generation sequencing using the Student’s t statistic as a phenotype for a double haploid (DH) population with one exceedingly “Yummy” parent, and one “yucky” parent. A check variety was used to evaluate each DH. Both “Yummy” and “yucky” parents, Louise and Yumai34, respectively, are soft white spring wheats. Genotyping-by-sequencing (GBS) was used to identify 655 single nucleotide polymorphic (SNP) markers, which were then used to create a linkage map with 24 linkage groups. Single marker-trait association was performed using the Student’s t phenotype for each marker, as well as composite interval mapping to identify quantitative trait loci (QTL). Thirty-nine SNP markers exhibited a significant marker-trait association with the Student’s t statistic. Additionally, four QTL were detected. In addition to the SNP markers, dominant markers with approximately 45-60% “missing” calls were evaluated, including both Louise-dominant and Yumai34-dominant markers. The Louise-dominant markers exhibited high levels of marker-trait association with the Student’s t statistic. Furthermore, many of these dominant markers were assigned to the same chromosome arms, giving seven groups of markers located on the same chromosome arms. In particular, dominant markers on 4DL exhibited the greatest F-values and suggested the strongest evidence of genes involved in wheat flavor located there. This work furthers the understanding of genes involved in “Yummy” and “yucky” flavor and will facilitate the development of varieties with greater consumer acceptance for whole-wheat products.