Classifying Oryza sativa accessions into Indica and Japonica using logistic regression model with phenotypic data

In Oryza sativa , indica and japonica are pivotal subpopulations, and other subpopulations such as aus and aromatic are considered to be derived from indica or japonica . In this regard, Oryza sativa accessions are frequently viewed from the indica/japonica perspective. This study introduces a computational method for indica/japonica classification by applying phenotypic variables to the logistic regression model (LRM). The population used in this study included 413 Oryza sativa accessions, of which 280 accessions were indica or japonica . Out of 24 phenotypic variables, a set of seven phenotypic variables was identified to collectively generate the fully accurate indica/japonica separation power of the LRM. The resulting parameters were used to define the customized LRM. Given the 280 indica/japonica accessions, the classification accuracy of the customized LRM along with the set of seven phenotypic variables was estimated by 100 iterations of ten-fold cross-validations. As a result, the classification accuracy of 100% was achieved. This suggests that the LRM can be an effective tool to analyze the indica/japonica classification with phenotypic variables in Oryza sativa .