准确地划分玉米杂种优势群, 可有效地指导种质的改良、 杂种优势模式的构建和新品种的选育。 目前主要利用系谱法、 配合力测定、 同工酶和分子标记等进行玉米杂种优势类群的划分, 这些划分方法成本高, 操作复杂, 甚至有的方法还需破坏种子。 本文探索了利用近红外光谱技术(NIRS)快速、 无损地划分玉米杂种优势群的可行性。 使用MPA型傅里叶变换近红外光谱仪漫反射模式采集整粒玉米种子的光谱, 用滑动平均窗口平滑、 一阶差分导数、 矢量归一化对光谱进行预处理。 将光谱用主成分分析方法降维到低维空间, 使用层次聚类法对A组6个中国的骨干自交系和B组的6个优良自选系共12个材料进行杂种优势群划分, 结果将A组6个自交系划分为3类, A1(郑58)和A2(掖478)为一类, A3(昌7-2)和A4(黄早四)为一类, A5(Mo17)和A6(四F1)为一类, 这与系谱法分析结果高度吻合。 对B组6个自选系, 将B1和B2, B3和B4, B5和B6分别划分为一类, 与SSR标记的类群划分结果一致。 由此可见, 利用NIRS划分玉米杂种优势类群是简便、 快速、 高效、 可行的方法。

Accurate division of heterotic group in maize can provide effective information on germplasm improvement, heterosis mode construction, and development of new varieties. The main methods used at present to divide maize heterotic group are pedigree, combining ability test, isozyme markers, and molecular markers. These methods are inferior because of their high cost and operational complexity. Some of them even have to destroy the seeds. This study explored novel method feasibility of near infrared (NIR) spectroscopy in quick nondestructive heterotic grouping of maize. The near infrared diffuse reflection spectra of maize seeds were collected and preprocessed using smoothing, first derivative, and vector normalization. The features of these spectra were extracted through principal component analysis (PCA). Twelve Chinese maize inbred line samples were selected, including six leading inbred lines (Part A) and six excellent self-selection lines (Part B). Part A was divided into three groups with NIR spectroscopy, namely, A1 (Zheng58) and A2 (Ye478), A3 (Chang7-2) and A4 (Huangzaosi), and A5 (Mo17) and A6 (SiF1). This division was in accordance with pedigree analysis. Part B was also divided into three groups by NIR spectroscopy: B1 and B2, B3 and B4, as well as B5 and B6. This division conforms to clustering analysis based on SSR molecular marker. These processes confirm that NIR spectroscopy is a convenient, highly efficient, and feasible heterotic grouping method of maize.