QTL mapping for 11 agronomic traits based on a genome-wide Bin-map in a large F2 population of foxtail millet (Setaria italica (L.) P. Beauv)

To identify quantitative trait loci (QTLs), a large F-2 population in foxtail millet including 543 plants from a cross between Aininghuang and Jingu 21 was used to construct a high-density linkage map based on restriction site-associated DNA sequencing (RAD-seq). Based on the map, QTLs for 11 various agronomic traits, consisting of PH, PL, PD, PNL, FID, SID, PW, GW, TGW, PT and PC, were mapped. The map contained 3129 Bin markers from 48,790 single nucleotide polymorphisms (SNPs) spanning 1460.996cM. Fifty-seven QTLs related to these 11 agronomic traits were detected using composite interval mapping (CIM), which explained 0.38-30.52% of total phenotypic variation explained (PVE). Among those, five major QTLs with a large PVE of more than 10%, including qPH5-2 for PH, qPD5-2 for PD, qPW5-1 for PW and qPC7-1 and qPC7-2 for PC, were detected. Notably, an extremely large effect QTL with a PVE of 30.52%, the qPH5-2 for PH, was observed. Furthermore, we newly developed five markers, which could be used for marker-assisted selection. Then, we verified these QTLs including qPC7-2 for PC, qPW5-1 for PW and qPH5-2 for PH were positive via correlation analysis of markers to traits in the natural population and the advanced generation population (AJF(5)) derived from the mapping population by single seed descent. Additionally, we found two multi-effect Mqtls, Mqtl5-2 and Mqtl2-5, underlying panicle development and yield, and one Mqtl, Mqtl1-2, probably related to plant growth and development by QTL prioritisation via Meta-QTL analysis. Moreover, we discussed candidate genes for the five major QTLs. Thus, mapping QTLs by RAD-seq using a large F-2 population is efficient, and the present study offers valuable insights into the genetic basis of quantitative traits and marker-assisted selection in foxtail millet.