Construction of a core collection of Perilla frutescens (L.) Britton Germplasm in the South Korean gene bank using agro-morphological traits

Perilla frutescens var. frutescens (PF) is widely cultivated in South Korea as an oilseed and a vegetable crop. Additionally, 2 weedy types, P. frutescens var. frutescens (WPF) and P. frutescens var. crispa (WPC), are commonly found in the country. This study aims to develop a core collection using 7 qualitative and 10 quantitative traits from 1227 Perilla accessions representing these 3 types (PF, WPF, WPC). Significant variations were observed among these types, with WPC displaying the most diverse leaf, stem, and flower colors, while PF had the largest plant size and seed weight. WPC exhibited the longest inflorescences but had lower values for most other traits. Several core collections were created: QL_Core (qualitative traits), QN_Core (quantitative traits), and A_Core (all traits + passport data), each containing 184 accessions (∼15 %). P_Core (PowerCore software) was the largest, with 362 accessions (∼29.5 %). The D_Core, consisting of 235 accessions (∼19.2 %), exhibited the highest variance and coefficient of variation, indicating the greatest diversity. Comparative analysis revealed that D_Core showed greater variation across all quantitative traits than the entire collection. The principal component analysis confirmed its genetic diversity, with a higher cumulative variation (90 % by PC5) than the entire collection (74.6 %). Boxplot analysis demonstrated wider dispersion, while analysis of variance indicated no significant differences in 7 traits. However, plant height, node number, and branch number were slightly higher in the entire collection. Overall, D_Core effectively represents the variation of Perilla accessions, making it a valuable collection for future genetic and breeding research. • Perilla frutescens var . frutescens is widely cultivated in South Korea as both an oilseed and a vegetable . • The three types of Perilla crop are clearly distinguished by their morphological characteristics. • To establish a core collection, 17 agro-morphological traits were evaluated across 1227 accessions of the three Perilla types. • D_Core effectively captures the diversity of the entire collection with greater variability. • The results of PCA, boxplot, and ANOVA confirm that D_Core serves as a robust representation of the entire collection.