Submitted to: Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: August 19, 1996
Publication Date: N/A
Interpretive Summary: Some phenotypic traits are controlled by one or few genes and segregate into specific expression classes, such as tall versus short plants. It is relatively easy to improve crops for such phenotypic traits using traditional breeding methods based on visual selection. However, many traits of agronomic interest are known as quantitative traits. They cannot be separated into discrete classes but exhibit a continuous range of expression making it difficult for a breeder to distinguish plants with "good" genes from those carrying "bad" genes. The genetic basis of quantitative traits is not fully understood, but several theories abound. For instance, if a trait is controlled by many genes, each with small effect, the phenotypic differences between the many possible genotypes would be minute, making it impossible to distinguish discrete classes. Alternatively, a quantitative trait may be controlled primarily by one or few large genes whose expression is modified by other genes. Such interaction between genes is known as epistasis. One example of epistasis would be when the enzymatic protein made by one gene is modified by a second gene so that it is more or less active, thus causing increased or decreased phenotypic expression of the first gene. Using molecular gene mapping techniques, we investigated the genetic control of several quantitative traits and concluded that epistasis does indeed play a significant role in the expression of quantitative traits. This information will be used by breeders and geneticists to improve breeding strategies for such quantitative traits.
The genetic basis for three grain yield components of rice - 1000 kernel weight (KW), grain number per panicle (GN), and grain weight per panicle (GWP) was investigated using 115 RFLP markers and F4 progeny testing from a cross between Lemont and Teqing. Following identification of 19 QTLs affecting these traits, we investigated the role of epistasis in genetic control of these phenotypes. Using both two-way ANOVA and multiple regression analyses, we were able to identify 63 markers throughout the genome which were involved in 79 highly significant interactions affecting the three traits. The majority of the putative epistatic loci did not appear to have 'main' effects on any of the traits, but influenced the three traits in a predominantly complementary manner. The more complex traits GN and GWP appeared to be more greatly influenced by epistasis than the highly heritable trait KW. These results indicate that epistasis is an important genetic basis for complex traits such as yield components, especially GWP and GN. The identification of epistatic loci is a step toward resolution of discrepancies between QTL mapping and classical genetic dogma, contributes to better understanding of the persistence of genetic variation for quantitative traits in populations, and impels reconsideration of optimal mapping methodology and marker-assisted breeding strategies for improvement of complex traits.