ASSOCIATION MAPPING OF DIFFERENT AGRONOMIC TRAITS IN RICE (Oryza sativa L.) GERMPLASM

Abstract

Rice in cereal group is the important food source for more than half of the world population. It has very important position among the cash crops of Pakistan. The country exports high quality rice to different world countries. The present study was carried out to identify loci related to different rice agronomic traits by association mapping technique. For this, a panel of 79 genotypes of Pakistan, Philippine, India, China and Malaysia was evaluated consecutively for three years from 2009-2011. Phenotypic data of different agronomic traits of these genotypes were collected at the adult plant stage. The ANOVA showed that the 79 rice germplasm are differed significantly from each other with respect to all the morphological traits studied. The phenotypic correlation coefficient among all rice agronomic traits was evaluated. The results showed that some of the parameters presented positive correlation, while other exhibited negative correlation with one another. The results of correlation exhibited that selection of higher number of tillers per plant, panicle length, panicle weight, thousand grain weight, number of grains per panicle, spikelet fertility and medium plant height would mainly encourage the rice breeders to achieve more paddy yield in rice. Development of semi dwarf variety will also improve rice yield. So these rice agronomic traits earn more consideration in future rice breeding programs for developing fine grain and higher yielding rice. Population structure of all 79 accessions was determined by using SSR (259) and SNP (733) markers separately and by combing both SSR+SNP markers (992). STRUCTURE software was used to visualize graphically the correct number of sub populations for association mapping study. Recently SNPs are being used frequently for population structure analysis. In present study by using 733 SNP markers on the population, the core collection was divided into two main groups G1 (japonica and basmati) and G2 (indica and other cultivars). G1 included Pakistani basmati varieties e.g. Basmati-370, Super Basmati and Basmati pak while Basmati-2000 showed above 80% similarity to japonica basmati group. Comparison of SSR and SNP result for population structure analysis showed that in case of SSR two at k=2 and four subgroups at k=4 were obtained while in xix SNP only two subgroups at k=2 were found. It may be due to the reason that SSR are multi-allelic markers while SNPs are bi-allelic in nature. So for population structure analysis SSR markers are more informative as compared to SNP markers. Two Statistical approaches such as GLM (general linear model) & MLM (mixed linear model) were applied for association study by using both SNP and SSR markers on diverse rice germplasm different rice parameters. Results illustrated that two different models identify rather different numbers of markers that were strongly associated with the parameters under study. In the GLM approach only Q matrix was used while in MLM both Q matrix and K matrix were used to accurate false positive (superior association). So, the number of markers strongly associated with the traits was much lower in MLM as compared to GLM. Only QTLs above LOD 3 were selected both in GLM and MLM. In these results mostly lower P values were observed in GLM as compared to those from MLM. The phenotypic data of rice agronomic traits such as plant height, number of tillers per plant, first leaf length, first leaf width, first leaf area, flag leaf length, flag leaf width, flag leaf area, panicle length, panicle weight, thousand grain weight, number of grains per panicle, spikelet fertility, paddy grain length, paddy grain width, polished grain length, polished grain width, polished grain shape, cooked grain length, cooked grain width cooked grain shape and yield per hectare were measured. Several markers appeared to be significantly linked to different rice parameters under study. Some of these markers were already reported in previous marker trait association study while many novel markers were also identified. This study has highlighted a unique set of loci which are associated with different rice agronomic traits and can be used for marker assisted breeding to enhance the rice performance under unfavorable conditions. Marker trait associations for number of tillers per plant, panicle length, panicle weight, thousand grain weight, number of grains per panicle, spikelet fertility and plant height which are found to be associated with yield in this study or in previous studies could also improve our understanding about the performance of genotypes.