Diversity analyses of improved rice (Oryza sativa L.) varieties and their progenitors using morphological characters and molecular markers

Morphology-based analysis showed that 41 traits contributed to the total variation in ancestral lines with Euclidean distances ranging from 3.970 to 17.389 while only 33 traits contributed to the total variation in improved varieties with Euclidean distances ranging from 2.228 to 16.706. Data showed that breeders selected for uniform traits, thus lowering the phenotypic divergence in improved varieties. Basal leaf sheath color, blade color, collar color, internode color, apiculus color, stigma color, culm length and plant height were the most important sources of variation in released varieties. On the otherhand, parental separation was strongly affected by variation in quantitative characters like plant height, culm length, heading, maturity, leaf length and productive tillers. Cluster analysis permitted separation of improved varieties into 2 major clusters; cluster 1 consisted of 72 (92 percent) varieties having green-colored plant parts and cluster 2 consisted of 6(8 percent) varieties (BPI76); (PSB Rc34, IR46, UPL Ri3, UPL Ri5, C168) having purple-colored plant parts. Results of cluster analysis in ancestral lines allowed separation of O. nivara to the rest of the O. sativa group. In molecular-based analysis, 20 aligonycleotide RAPD primers and 4 microsatellite primer pairs produced 119 polymorphic bands in ancestral lines with Nei's genetic distances ranging from .044 to .920 and only 80 polymorphic bands in improved varieties with Nei's genetic distances ranging from .035 to .930. Data showed that rare alleles of the parentals were eliminated in the improved varieties as a consequence of modern plant breeding and an increase in monomorphic loci from ancestral lines to cultivated descendants, suggesting a low level of molecular diversity in the improved cultivars. IR70 formed singled cluster and diverged significantly from other improved varieties as shown by cluster analysis. The groupings of released varieties were based on the number of shared parentals, one served as a parent to another or they shared similar immediate paternal or maternal parent. Cluster analysis in ancestral lines allowed separation of O. nivara and BJ1 accessions from the other ancestral lines. Correspondence existed between geographical distribution and some clustering in the dendrogram. In overall clustering, morphology did not accurately group the improved varieties to known pedigree relationships. Only the molecular genetic distance showed good agreement with pedigree distances (r=0.26** ). Grouping of sister lines was evident in molecular rather than morphological method of analysis. Because of high degree of polymorphism detected by each primer, the simplicity and rapidity to obtain results, these PCR-based molecular markers may be preferred to determine diversity of improved varieties, especially on genotypes with unknown ancestry