Screening of rice drought-tolerant lines by introducing a new composite selection index and competitive with multivariate methods

Abstract

AbstractSelection and breeding for drought tolerance in rice have always been one of the leading objectives for rice breeders in water-deficient farming areas. In the present study, we applied the potential of recombinant inbred lines (RILs) population, which were derived from cross Shahpasand (Iranian landrace) and IR28, for the development of drought-tolerant rice lines. One hundred fifty-two lines along with five check varieties were investigated from 2017 to 2019 under non-stress and drought stress conditions. The yield reduction caused by drought based on overall mean during 2017, 2018, and 2019 were estimated to be 89.40, 57.95, and 35.31%, respectively. Using different statistical methods, certain lines, including L33, L90, and L109, which are considered as the best lines in most environments, were found to be promising for being utilized to increase rice drought tolerance. The averages of grain yield of the above-mentioned lines were respectively 6.45, 5.80, and 5.70 t ha−1 under non-stress condition, and respectively 2.77, 2.66, and 2.59 t ha−1 under drought stress condition. The yield reduction of the selected lines were significantly lower than that of others indicating the significant transgressive segregation. The results revealed using the combination of the best identified tolerance and susceptibility indices and GT-biplot are effective methods for screening superior lines. However, their utilization is not easy and requires specialized packages. For the first time, we introduced a new composite index as a combination of significant indices (CSI). CSI is in the form of a linear function of indices which effectiveness is determined by their correlation coefficient with grain yield. According to our results, using CSI, the identified rice drought-tolerant lines were in high agreement with those obtained by other methods, demonstrating that CSI is a simple but reliable composite index.