Identification of molecular markers associated with sugar-related traits in a Saccharum interspecific cross

Abstract

The cultivated sugarcane (Saccharum spp. hybrids, 2n = 100–130) is one crop for which interspecific hybridization involving wild germplasm has provided a major breakthrough in its improvement. Few clones were used in the initial hybridization event leading to a narrow genetic base for continued cultivar development. Molecular breeding would facilitate the identification and introgression of novel alleles/genes from the wild germplasm into cultivated sugarcane. We report the identification of molecular markers associated with sugar-related traits using an F1 population derived from a cross between S. officinarum ‘Louisiana Striped’ × S. spontaneum ‘SES 147B’, the two major progenitor species of cultivated sugarcane. Genetic linkage maps of the S. officinarum and S. spontaneum parents were produced using the AFLP, SRAP and TRAP molecular marker techniques. The mapping population was evaluated for sugar-related traits namely, Brix (B) and pol (P) at the early (E) and late (L) plant growing season in the plant cane (04) and first ratoon (05) crops (04EB, 04LB, 04LP, 05EB and 05EP). For S. officinarum, combined across all the traits, a total of 30 putative QTLs was observed with LOD scores ranging from 2.51 to 7.48. The phenotypic variation (adj. R2) explained by all QTLs per trait ranged from 22.1% (04LP) to 48.4% (04EB). For S. spontaneum, a total of 11 putative QTLs was observed with LOD scores ranging from 2.62 to 4.70 and adj. R2 ranging from 9.3% (04LP) to 43.0% (04LB). Nine digenic interactions (iQTL) were observed in S. officinarum whereas only three were observed in S. spontaneum. About half of the QTLs contributed by both progenitor species were associated with effects on the trait that was contrary to expectations based on the phenotype of the parent contributing the allele. Quantitative trait loci and their associated effects were consistent across crop-years and growing seasons with very few QTLs being unique to the early season. When the data were reanalyzed using the non-parametric discriminant analysis (DA) approach, significant marker-trait associations were detected for markers that were either identical to or in the vicinity of markers previously identified using the traditional QTL approach. Discriminant analysis also pointed to previously unidentified markers some of which remained unlinked on the map. These preliminary results suggest that DA could be used as a complementary approach to traditional QTL analysis in a crop like sugarcane for which saturated linkage maps are unavailable or difficult to obtain.