AbstractBackgroundThe structural reinforcement of cell walls by hydroxycinnamates has a significant role in defense against pests and pathogens, but it also interferes with forage digestibility and biofuel production. Elucidation of maize genetic variations that contribute to variation for stem hydroxycinnamate content could simplify breeding for cell wall strengthening by using markers linked to the most favorable genetic variants in marker-assisted selection or genomic selection approaches​.ResultsA genome-wide association study was conducted using a subset of 282 inbred lines from a maize diversity panel to identify single nucleotide polymorphisms (SNPs) associated with stem cell wall hydroxycinnamate content. A total of 5, 8, and 2 SNPs were identified as significantly associated top-coumarate, ferulate, and total diferulate concentrations, respectively in the maize pith. Attending to particular diferulate isomers, 3, 6, 1 and 2 SNPs were related to 8–O–4 diferulate, 5–5 diferulate, 8–5 diferulate and 8–5 linear diferulate contents, respectively. This study has the advantage of being done with direct biochemical determinations instead of using estimates based on Near-infrared spectroscopy (NIRS) predictions. In addition, novel genomic regions involved in hydroxycinnamate content were found, such as those in bins 1.06 (for FA), 4.01 (for PCA and FA), 5.04 (for FA), 8.05 (for PCA), and 10.03 and 3.06 (for DFAT and some dimers).ConclusionsThe effect of individual SNPs significantly associated with stem hydroxycinnamate content was low, explaining a low percentage of total phenotypic variability (7 to 10%). Nevertheless, we spotlighted new genomic regions associated with the accumulation of cell-wall-bound hydroxycinnamic acids in the maize stem, and genes involved in cell wall modulation in response to biotic and abiotic stresses have been proposed as candidate genes for those quantitative trait loci (QTL). In addition, we cannot rule out that uncharacterized genes linked to significant SNPs could be implicated in dimer formation and arobinoxylan feruloylation because genes involved in those processes have been poorly characterized. Overall, genomic selection considering markers distributed throughout the whole genome seems to be a more appropriate breeding strategy than marker-assisted selection focused in markers linked to QTL.