|
Two approaches of genetic analysis of quantitative traits were compared with a case study on soybean. One approach was the segregation analysis developed by Gai et al. (2003), which utilized information from individuals of one or multiple segregation populations as well as that from parents based on the principles of the major-gene plus polygene inheritance model, mixture distribution, joint maximum-likelihood function, IECM (Iterated Expectation and Conditional Maximization) algorithm, and Akaike’s information criterion and goodness of fit tests. Another approach was quantitative trait locus (QTL) mapping with molecular markers. A recombinant inbred line (RIL) population with 201 families derived from (Kefeng No.1×1138-2) F2:7:10 along with their parents were tested in a randomized block design experiment. The 171 RFLP, 60 SSR, and 79 AFLP molecular markers were used to mark the 201 families. The data of nine traits, i.e., number of days to flowering, number of days to maturity, plant height, number of nodes on main stem, number of pods per node, 100-seed weight, protein content, oil content, and plot yield, were analyzed with the segregation analysis procedure of RIL population with parents (Gai et al., 2003; Zhang and Gai, 2000; Zhang et al., 2001) to detect their genetic system, and those along with the molecular marker data were analyzed with WinQTL Cartographer (Basten et al., 1999; Zeng, 1993, 1994) to detect their QTL system. The results showed that both procedures could detect the main major genes or QTLs, and therefore, could be used as a mutual check and supplement. From the results that most of the traits were mainly controlled by three or four QTLs, it was impressed that the segregation analysis procedure of four major-gene plus polygene mixed inheritance model should be developed to fit the requirements. The results also showed that the QTLs of the involved traits concentrated on several linkage groups, such as C2, B1, F1, M, and N. Finally, the results showed that the experimental sample was not necessarily coincident with the theoretical population according to equality test, symmetry test, and representation test, and therefore, the sample should be checked, tested and then adjusted to fit the theoretical requirements through deleting the extra-biased families and markers. |
