基于“LabelmeP1.0”提取的高粱维管束参数与农艺性状关联研究

Relationship of“LabelmeP1.0”-derived vascular parameters with agronomic traits in sorghum

茎秆维管束对作物机械支撑和物质运输具有重要作用。为了高通量导出作物茎秆切片的维管束参数,本研究开发了基于Labelme标注图片和Python编程语言的方法“LabelmeP1.0”。利用该方法导出了92份高粱种质资源的茎秆维管束和区域相关的26个参数。通过相关分析和主成分分析,从中归纳了各区域面积以及和维管束数量、大小、位置、比例和密度相关的15个参数作为核心指标。研究还发现,高粱维管束性状和农艺学性状关系密切,维管束数目、密度与穗重、粒重和一级枝梗数正相关。进一步可将资源材料分为3类,分别为矮粗薄皮、种子少和维管束大而多的类Ⅰ,高个细茎、种子少和维管束少而稀的类Ⅱ,以及矮粗厚皮、种子多和维管束数多的类Ⅲ。本研究可为禾本科作物的茎秆解剖学结构参数研究提供技术和新思路。

Stem vascular bundle plays an important role in the mechanical supporting and material transportation of crops.To extract the vascular bundle parameters from the crop stem slices,we developed a high-throughput method“LabelmeP1.0”,which combined the image-annotation function of commercial software“Labelme”with the Python programming language.A total of 26 parameters including both the primary and secondary parameters related to stem regions and vascular bundles was generated in 92 sorghum germplasm resources.15 parameters related to the number,size,location,proportion,and the density of vascular bundle were summarized as the core indexes by correlation analysis and principal component analysis.In addition,the number and the density of vascular parameters were positively related to agronomic traits including panicle weight,grain weight,and number of primary rachis branch.The germplasm resources could be divided into three categories,namely,class I with lower plant height,thicker stem,thinner rind,fewer seeds,and more vascular bundles,in contrast class II with higher plant height,thinner stems,fewer seeds and vascular bundles and,class III with lower plant height,thicker stem,more seeds and vascular bundles.This study provides the methodology and new ideas for the study of the stem anatomical parameters in crops.