Rice Genetics IV - IRRI books - International Rice Research Institute

M-QTLsM-QTLs are defined as single Mendelian factors at which effects (additive and/ordominance) on a given phenotype arise from allelic substitution and are detected bymarker-trait associations using single-factor ANOVA or interval mapping models(Lander and Botstein 1989, Zeng 1994, Li 1997). M-QTLs in rice appear to includetwo groups of genes. The first group includes major genes of very large effectson highly heritable traits, which are typically detected with very large LOD scores(>10.0), and each explains a large portion of the total trait variation in a mappingpopulation. Examples of this type are sd-1 for semidwarf stature in rice, Xa4 forbacterial blight resistance, Ta9 for tiller angle, Hd-1, Hd-3, and QHd3 for headingdate, etc. (Table 1).The second group includes the typical M-QTLs, which represent most (more than90%) QTLs reported to date. These typical M-QTLs tend to have relatively smalleffects. There are two general results regarding the M-QTLs from previous studies.First, the number of detected M-QTLs for a specific trait in a population evaluated ina specific environment is relatively small. Based on the 324 cases (trait/population/environment combinations) in the previous rice QTL mapping studies involving 46mapping populations and 71 phenotypes, the average number of detected M-QTLsper trait/population/environment is 3.7 ± 1.2 and, surprisingly, this number does notdiffer between high- and low-heritability traits. For example, the average detectablenumber of M-QTLs per population/environment is 3.7 and 4.1 for two highly heritabletraits, days to heading and plant height, respectively, and 3.3 and 3.4 for the lowheritabilitytraits, grains per panicle and grain yield, respectively. However, when acomparison is made across mapping populations and environments, many more M-QTLs are detected for each trait, and these are widely distributed on the 12 rice chromosomes(Table 2). This underestimation of M-QTL number in most QTL mappingstudies is due largely to epistasis and QE interactions, which will be discussed later.Second, the accuracy of the estimated M-QTL genetic parameters, such as their effects,genomic locations, and gene actions, varies considerably, depending largely onerrors in phenotyping and the statistical methods for parameter estimation. In theTable 1. Some major genes affecting quantitative traits detected as main-effect QTLs in rice.M-QTL Trait Chromosome LOD Referencesd-1 Plant height 1 17.5 Huang et al (1996)Xa4 Bacterial blight resistance 11 66.5 Li et al (1999b)Sub1 Submergence tolerance 9 58.4 Sripongpangkul et al (2000), Xu andMackill (1995), Nandi et al (1997)Ta9 Tiller angle 9 32.3 Li et al (1999a)QIne1 Internode elongation 1 21.5 Sripongpangkul et al (2000)QHd3 Heading date 3 24.5 Li et al (1995a)Hd-1 Heading date 6 44.2 Yano et al (1997)Hd-3 Heading date 6 64.4 Yano et al (1997)QFll3b Flag leaf length 3 11.1 Li et al (1998)QSh2 Grain shattering 2 16.4 Zhong et al (1999), Fukuta et al(1996)QTL mapping in rice: . . . 155