Mapping and characterisation of QTL × E interactions for traits determining grain and stover yield in pearl millet

Abstract.

A mapping population of 104 F₃ lines of pearl millet, derived from a cross between two inbred lines H 77/833-2 × PRLT 2/89-33, was evaluated, as testcrosses on a common tester, for traits determining grain and stover yield in seven different field trials, distributed over 3 years and two seasons. The total genetic variation was partitioned into effects due to season (S), genotype (G), genotype × season interaction (G × S), and genotype × environment-within-season interaction [G × E(S)]. QTLs were determined for traits for their G, G × S, and G × E(S) effects, to assess the magnitude and the nature (cross over/non-crossover) of environmental interaction effects on individual QTLs. QTLs for some traits were associated with G effects only, while others were associated with the effects of both G and G × S and/or G, G × S and G × E(S) effects. The major G × S QTLs detected were for flowering time (on LG 4 and LG 6), and mapped to the same intervals as G × S QTLs for several other traits (including stover yield, harvest index, biomass yield and panicle number m^–2). All three QTLs detected for grain yield were unaffected by G × S interaction however. All three QTLs for stover yield (mapping on LG 2, LG 4 and LG 6) and one of the three QTLs for grain yield (mapping on LG 4) were also free of QTL × E(S) interactions. The grain yield QTLs that were affected by QTL × E(S) interactions (mapping on LG 2 and LG 6), appeared to be linked to parallel QTL × E(S) interactions of the QTLs for panicle number m^–2 on (LG 2) and of QTLs for both panicle number m^–2 and harvest index (LG 6). In general, QTL × E(S) interactions were more frequently observed for component traits of grain and stover yield, than for grain or stover yield per se.