Abstract:

Grain length (GL), grain width (GW), length width ratio (LWR) and grain weight(GWt) are most important apparent quality traits in rice, which are positively correlated with the commercial value of rice especially in international market. Moreover, grain shape is also a vital factor related to rice yield and milling quality. Companed with GL, GW and LWR, grain weight is one of the most important components related to rice yield. Several studies reported that GL, GW, LWR and GWt were controlled by some QTLs were detected under lowland condition. There have been no reports about QTLs for grain shape and grain weight under both upland and lowland conditions. Therefore, it is necessary to analyze the correlation and inheritance of grain shape and grain weight between under upland and under lowland conditions, especially to find QTLs under upland condition.
A doubled haploid (DH) population consisting of 116 DH lines developed through anther culture of the F1 hybrid between an upland rice cultivar ‘IRAT109’ and a lowland rice cultivar ‘Yuefu’ was used in this study. The parents and the 116 DH lines were planted under water- and dry-cultivation conditions, on the experiment farm of China Agricultural University in 2002. Seeds of parents and 116 DH lines were harvested at maturity, and used to measure grain length, grain width, length width ratio and grain weight. QTLs analysis was conducted with QTL Mapper version 1.0 and the constructed molecular linkage map consisting of 165 molecular markers, which covered 1 535 cM in length with an average distance of 9.3 cM between adjacent markers. A threshold probability of P<0.005 was used to declare the existence of additive QTLs. Correlation analysis between grain shape and grain weight was conducted using the software of SPSS.
Positively significant correlations between GL and LWR and between GL and GWt in both upland and lowland environments were observed. Negatively significant correlations between GW and LWR and positively significant correlations between GW and GWt under the two environments were also detected. This result indicated that GL, GW, LWR and GWt had closely genetic background. All the four traits in upland environment were significantly correlated with those in lowland environment respectively, with the biggest correlation coefficient of 0.817 for GL and the smallest coefficient of 0.457 for GW. Therefore, compared to GW, GL was less affected by water stress (upland condition).
A total of 14 QTLs were detected for GL, GW, LWR and GWt, which located on chromosome 1, 5, 6, 7,10 and 12. Ten QTLs had high general contribution with explained variance of over 10%. For GL, one QTLs under upland condition and four QTLs under lowland condition were detected respectively. These QTLs explained 5.97%–28.85% of phenotypic variance. The LOD score was 1.93–5.11. For GW, only one QTLs was detected under upland condition. The LOD score and general contribution was 2.39 and 12.76% respectively. For LWR, one QTLs under upland condition and two QTLs under lowland condition was detected respectively. These QTLs explained 7.78%–21.89% of phenotypic variance. The LOD score was 2.08–4.60. For GWt, there were two additive QTLs under upland condition and three additive QTLs under lowland condition respectively. They explained 4.1%–14.8% of phenotypic variance. The LOD score was 2.68–9.45. Three QTLs qGL-5, qGWt-1a and qGWt-1b were both detected under upland and lowland environments (common QTL). Two co-localized QTLs (qGL-1a and qGWt-1a, and qGL-1b and qGWt-1b) controlling GL and GW were detected. These QTLs would be helpful in MAS of grain shape breeding.

Key words: Rice, Grain shape, Grain weight, Doubled haploid population, Quantitative trait loci