Genomic Prediction of Biological Shape: Elliptic Fourier Analysis and Kernel Partial Least Squares (PLS) Regression Applied to Grain Shape Prediction in Rice (Oryza sativa L.)

• DOI: 10.1371/journal.pone.0120610
• Publication Date: 2015-04-01T14:15:52Z
• Authors: Hiroyoshi Iwata; Kaworu Ebana; Yusaku Uga; Takeshi Hayashi
• Keywords: Kernel (algebra); Cross-validation

Shape is an important morphological characteristic both in animals and plants. In the present study, we examined a method for predicting biological contour shapes based on genome-wide marker polymorphisms. The expected to contribute acceleration of genetic improvement shape via genomic selection. Grain variation observed rice (Oryza sativa L.) germplasms was delineated using elliptic Fourier descriptors (EFDs), predicted single nucleotide polymorphism (SNP) genotypes. We applied four methods including kernel PLS (KPLS) regression building prediction model grain shape, compared accuracy cross-validation. analyzed multiple datasets that differed density sample size. Datasets with larger size higher showed accuracy. Among methods, KPLS highest Although ridge (RR) had equivalent dataset, result suggested potential high-dimensional EFDs. Ordinary PLS, however, less accurate than RR all datasets, suggesting use non-linear necessary method. Rice can be accurately SNP proposed useful selection shape.