目的 为筛选川贝母Fritillaria cirrhosa优质种质资源，开展新品种和良种选育，通过性状变异进行表型调查。方法 采用隶属函数和偏最小二乘-回归分析（partial least square-regression analysis，PLS-R）、相关性分析（correlation analysis，CA）、聚类分析、主成分分析（principal component analysis，PCA）等方法，对23份种质资源的10个农艺性状、5个生理和9个品质性状进行综合评价。结果 农艺性状的多样性指数为1.808～2.337，其中鳞茎鲜质量和果质量的多样性指数最高，均为2.30。24个数量性状的变异系数为6.63%～102.36%，其中农艺性状的变异主要来源于果质量、花朵数目和鳞茎鲜质量；品质指标的变异主要来源于鸟苷、腺苷和胸苷的差异。相关性分析显示生物碱含量和鳞茎鲜质量与茎粗、叶长、果长、果直径以及果质量等都表现出了正相关关系。聚类分析将收集的种质分成3大类群，第Ⅰ类群具有鳞茎鲜质量、生物碱含量和果实质量较高的特点，第Ⅲ类群中核苷类物质含量较高。PCA将24个性状简化为6个主成分，累积贡献率为78.746%。PLS-R与偏最小二乘-判别分析（partial least square-discriminant analysis，PLS-DA）揭示果直径、叶宽和脱氧腺苷、鸟嘌呤对D值的贡献较大，可作为筛选优良种源的关键指标。结论 综合排名前3位的种质为ZA、HG、GH表型，具有品质优，含量丰富的特点，可作为潜在的良种进行培育。川贝母农艺性状中的叶宽、茎粗和鳞茎鲜质量，以及品质性状中的脱氧腺苷、鸟嘌呤和尿嘧啶，是区分川贝母优劣的关键性状。对川贝母表型进行了较为全面的分类并开展资源系统评价，可为川贝母遗传资源挖掘和良种选育奠定基础。

Objective In order to screen high-quality germplasm resources of Fritillaria cirrhosa and conduct selection and breeding of new and improved varieties, this study conducted a phenotypic survey through trait variations.Methods Various methods were employed, including affiliation function, partial least squares-regression analysis (PLS-R), correlation analysis (CA), cluster analysis, and principal component analysis (PCA), to comprehensively evaluate 10 agronomic, five physiological, and nine qualitative traits among 23 germplasm resources. Results The results demonstrated that the diversity indices of agronomic traits ranged from 1.808 to 2.337, with the highest diversity indices for bulb fresh weight and fruit weight, both of which were 2.30. The coefficients of variation for the 24 quantitative traits ranged from 6.63% to 102.36%, with the variations of the agronomic traits mainly originating from the variations in fruit weight, number of flowers, and bulb fresh weight; and those of the quality indexes mainly originating from the differences in guanosine, adenosine and thymosine. Correlation analyses revealed positive associations between alkaloid content and bulb fresh weight with stem thickness, leaf length, fruit length, fruit diameter, and fruit weight. Cluster analysis classified the collected germplasm into three major groups. Group I was characterised by higher bulb fresh weight, alkaloid content and fruit weight, and group III had a higher content of nucleosides. PCA simplified 24 traits into six principal components with a cumulative contribution of 78.746%. PLS-R and partial least square-discriminant analysis (PLS-DA) revealed that fruit diameter, leaf width and deoxyadenosine and guanine contributed more to the D value, which can be used as a key index to screen for good seed sources.Conclusion The top 3 comprehensively ranked germplasm were ZA, HG, and GH phenotypes, which were characterised by excellent quality and rich content. These varieties hold potential as valuable seed sources. Leaf width, stem thickness and bulb fresh weight in agronomic traits, and deoxyadenosine, guanine and uracil in quality traits were screened as the key traits to differentiate the superiority and inferiority of F. cirrhosa. The present study provides a more comprehensive classification and resource systematic evaluation of the phenotypes of F. cirrhosa, serving as a foundation for genetic resource exploration and breed selection of F. cirrhosa.

R286.2

中国中医科学院科技创新工程项目（CI2024C003YN，CI2021A03910）;国家重大新药创制科技重大专项（2019ZX09201005-006-001）;中央级公益性科研院所基本科研业务费专项资金（L2023022，Z2023014，ZXKT22061，ZZ18-YQ-036）