45. Correlation coefficient analysis in identifying molecular markers associated with seedling rice root traits
  J. XU, L. ZOU and L. ZHU

Institute of Genetics, Chinese Academy of Sciences Beijing 100101, China

Most important agronomic characters are quantitative traits controlled by many major and minor genes. Rice varieties can differ markedly in the quantitative trait loci (QTL) that condition each trait. For systematic dissection of quantitative traits, mapping populations are essential. In this report, different rice varieties are evaluated for root traits and new correlated genetic markers are identified by correlation coefficient analysis.

Eight-four rice varieties were selected and planted in a hydroponics system (Xu et al., 2001). After 10 days growth in a nutrient solution, maximum root length (MRL) and root dry weight (RDW) were evaluated and characterized for each variety. The score for MRL was between 3.63~16 cm in an average of 9.28 cm while for RDW was between 9.75~33.7 mg in an average of 20 mg. Both traits show normal distribution in 84 rice varieties.

According to the root trait data, 27 of the rice varieties with extreme value were selected representing for the AFLP analysis. Genome DNA was isolated from each rice variety leaves and digested with restriction enzyme of Pst 1 and Taq 1 After ligation, pre-amplification and selective amplification, the products were separated on 6% of polyacrylamide gel. Each polymorphic band was scored and input into the database. The variety with the polymorphic band was labeled as genotype "1" while no polymorphic band was labeled as genotype "0". Correlation coefficient (r) between marker genotype and trait phenotype was considered as the parameter for the correlated marker determination. The formula was as follow:

where lxy was the covariance between trait phenotype and marker genotype; lxx and lyy separately were the variance of trait and marker. A positive "r" indicates the band facilitated trait expression, while a negative "r" implies inhibition.

Fifteen pairs of AFLP primer combinations were used for the correlated marker survey. At a significance level of 1%, 7 bands were totally identified from 4 primer combinations (Taq3/Pst3, Taq10/Pst3, Taq10/Pst5, Taq4/Pst4). One marker identified by Taq3/Pst3 was found correlated with both MRL and RDW, with correlation coefficients of -0.471 and -0.500, respectively. After cloning and sequencing, a pair of specific primers for this fragment was designed as Z336:


Figure 1 showed the amplification result by Z336 in 27 rice varieties ranging by the value

of MRL from large to small. Obviously, the polymorphic band distributed biased. Further analyzing the total 84 rice varieties, the correlation coefficient for MRL was computed as -0.193, which closed to the significant level of 5% while for RDW was -0.391 that was over the significant level of 1%. We conclude that Z336 is a QTL with negative control for the RDW expression. It could explain the RDW variation of 15.3%.

For anchoring the marker on rice chromosome, a molecular linkage map from a double haploid population of ZYQ8/JX17 was conducted. By linkage analysis, the marker Z336 was found linked to marker L190 at a distance of 9.4 cM on rice chromosome 11 (Fig. 2).


Xu, J., J.Li, X. Zhang, L. Zou, and L. Zhu, 2001, QTL Mapping of the Seedling Rice Root Traits. Acta Genetica Sinica, 28: 433-438.