To study the feasibility of ultilizing advanced generations of populations from the same cross for detecting QTL for yield and yield components, F₂ and F₇ populations were obtained from the cross Zhenshan 97B/Milyang 46. One-hundred and seventy one F₂ plants and 231 recombinant inbred lines (RIL) were grown in 1995 and 1999, respectively. For each RIL, two replications with 12 plants per replication were grown, and the middle eight plants were sampled. A single plant of each RIL was grown and sampled as well. Six traits, grain yield (GYD), number of panicles (NP), number of filled grains per panicle (NFGP), total number of spikelets per panicle (TNSP), spikelet fertility (SF) and 1000grain weight (TGWT), were scored.

Using data from the F₂ and RIL populations, linkage maps consisting of 109 and 115 RFLP markers were constructed, respectively. A putative QTL was detected using the threshold of LOD>2.0 for MAPMAKER/QTL 1.1b based on interval mapping (IM method; Lincoln et al. 1992), and using that of P<0.005 for QTLMAPPER 1.0 based on mixed linear model (MCIM method; Wang et al. 1999). For the additive and dominance

effects of a given QTL, only the former could be inferred using RIL alone. Therefore, the QTL detected in the F₂ population, only those having an additive effect higher than its dominance effect are discussed in this report. Such QTL are hereafter referred to as additive QTL.

1. Additive QTL detected in the F₂ population

Fourteen additive QTL were detected in the F₂ population, of which 8 were detected

with both methods (Table 1). Those QTL detected with only one method generally had lower LOD scores. For example, of the 5 QTL for TNSP detected by using IM method, only the QTL having the lowest LOD score was not detected by using MCIM method.

2. QTL detected in the RIL population

Two sets of data were collected for the RIL population. One set was the average value of 2 x 8 plants, the other was the value of single plants. Combining two mapping methods, four sets of QTL were detected. Irrespective of IM or MCIM method employed, QTL of higher LOD score detected based on the replicated trial, were always detected based on the single-plant-based trial. For example, of the QTL detected based on the replicated trial by using IM method, those detected with LOD score higher than 3.0 were always detected based on single-plant-based trial by using the same method (Table 2).

3. Comparison of QTL detection in the F₂ and RIL populations

Of the 14 additive QTL detected in the F₂ population, 10 were detected in the RIL population. In addition, QTL having the highest LOD score among those conditioning the same trait were always detected in the RIL population.

Generally, QTL detected with higher LOD scores in the RIL population were also detected in the F₂ population. For example, of the QTL detected based on the average values by using IM method, those having the highest LOD score detected for each of NFGP, TNSP, TGWT and GYD were always detected in the F₂. In any intervals where two or more QTL were detected, the QTL having the highest LOD score was always detected in the F₂. Moreover, when a QTL was detected in both generations, the direction of the additive effect always remained unchanged.

References

Lincoln S., M. Daley and E. Lander, 1992. Mapping genes controlling quantitative traits with MAPMAKER/ QTL1.1.Whitehead Institute Technical Report, 3rd edition, Whitehead Institute, Cambrige, Mass.

Wang D.L., J. Zhu, Z.K. Li and A.H. Paterson, 1999. Mapping QTLs with epistatic effects and QTL x environment interactions by mixed linear model approaches. Theor. Appl. Genet., 99: 1256-1264.