Quantifying the Spiral Leaf Trait of Arabidopsis from the 3D Shape Model Towards Computational Phenomics

Abstract

Computational phenomics becomes more important in the recently remarkable advance of functionalgenomics. In the RIKEN Genomic Sciences Center, more than 50,000 mutant lines of the modelplant Arabidopsis thaliana have been produced for the saturated mutagenesis of the genome[3]. Forthe research of functional genomics, to establish the high-throughput phenotypic screening systemis a matter of vital importance[1]. Towards the high-throughput system, we have proposed a novelmethodology of the phenotypic analysis which is based on 3-dimensional(3D) model reconstruction viasurface measurement by the laser range nder(LRF)[5] and the computational screens of morphologi-cal traits on the 3D reconstructed model[2]. However, conventional description of morphological traitsof Arabidopsis is mostly qualitative due to subjective human observations. Thus we need to de nequantitatively individual morphological traits to be handled as objectively digital parameters. In thisreport, we focus on the spiral leaf trait[3, 4] as one of 3D-speci c traits at rosette leaves, and proposea quantitative de nition of the spiral leaf trait. Through an experiment to extract the quantitativespiral leaf trait at a wild-type and a mutant line, we exhibit the importance for quantifying traitsprecisely towards computational screens based on the 3D reconstructed model.