Genome-wide association study of stem structural characteristics that extracted by a high-throughput phenotypic analysis "LabelmeP rice" in rice.

Abstract

Stem is important for assimilating transport and plant strength; however, less is known about the genetic basis of its structural characteristics. In this study, a high-throughput method, "LabelmeP rice" was developed to generate 14 traits related to stem regions and vascular bundles, which allows the establishment of a stem cross-section phenotype dataset containing anatomical information of 1738 images from hand-cut transections of stems collected from 387 rice germplasm accessions grown over two successive seasons. Then, the phenotypic diversity of the rice accessions was evaluated. Genome-wide association studies identified 94, 83, and 66 significant single nucleotide polymorphisms (SNPs) for the assayed traits in 2 years and their best linear unbiased estimates, respectively. These SNPs can be integrated into 29 quantitative trait loci (QTL), and 11 of them were common in 2 years, while correlated traits shared 19. In addition, 173 candidate genes were identified, and six located at significant SNPs were repeatedly detected and annotated with a potential function in stem development. By using three introgression lines (chromosome segment substitution lines), four of the 29 QTLs were validated. LOC_Os01g70200, located on the QTL uq1.4, is detected for the area of small vascular bundles (SVB) and the rate of large vascular bundles number to SVB number. Besides, the CRISPR/Cas9 editing approach has elucidated the function of the candidate gene LOC_Os06g46340 in stem development. In conclusion, the results present a time- and cost-effective method that provides convenience for extracting rice stem anatomical traits and the candidate genes/QTL, which would help improve rice.

Plant phenotyping relevance

イネ茎断面画像から解剖学的形質を抽出する高スループット手法を開発し、形質データセット構築と検証に用いており、表現型取得法が研究の中心である。

Code and data availability

The paper's phenotyping/analysis Python code (LabelmeP rice pipeline) is explicitly provided as authors' supporting information Data S1, publicly available with the online article. Other referenced URLs (SNP-Seek, Q-TARO, RiceRNA, rmbreeding) are external databases, not paper-specific assets.