Comprehensive phenomics and vegetative yield analysis of global kale (Brassica oleracea var. acephala) germplasm in controlled environment agriculture

TL;DR

This study analyzes the traits of 203 kale varieties under controlled agriculture conditions to identify key characteristics linked to higher yield.

Contribution

The study provides a comprehensive phenomics baseline and identifies key traits associated with kale yield in controlled environments.

Findings

203 kale accessions were analyzed for 113 traits, revealing significant genetic variability.

28 phenotypes were identified as most strongly associated with total aboveground fresh weight, a key yield indicator.

Three principal components explained 87.1% of yield variation, related to plant size, canopy structure, and density.

Abstract

Kale (Brassica oleracea var. acephala) is a high value leafy vegetable with an extensive domestication history and germplasm diversity, making it an ideal target for genetic improvement. To meet growing food security needs particularly with controlled environment agriculture (CEA) systems, specialized breeding strategies are required. The goal of this study was to survey the phenotypic architecture of a global kale germplasm collection under commercial CEA conditions. This study establishes a phenotypic baseline and serves as a hypothesis generating resource for future genetic and physiological studies in kale and other leafy vegetables grown under CEA.

A total of 203 kale accessions were phenotyped for 113 quantitative traits using high-throughput phenotyping methods. Significant differentiation was observed across all traits, with coefficient of variation ranging from 2.5% to 180.7%, confirming broad genetic variability among accessions. Trait correlation networks and hierarchical clustering grouped phenotypes into seven biologically corresponding modules including leaf, stem and root morphology, plant architecture, hyperspectral indices, and seedling growth. These modules highlight coordinated phenotypic patterns among traits. Integrative yield analyses combining partial least squares variable importance in projection with differential trait analysis identified 28 phenotypes most strongly associated with total aboveground fresh weight, a robust proxy for CEA vegetative yield. Principal component analysis further distilled these traits into three orthogonal components explaining 87.1% of total yield variation. These components represented modules related to plant organ size, canopy structure, and density, emphasizing their biological contribution to harvestable biomass.

This study generates a foundational phenomics resource and comprehensive dissection of kale’s yield architecture under CEA conditions. The composition of traits identified constitutes a targeted set of breeding traits to be further validated for improved leafy vegetable yield. By integrating large-scale germplasm resources with phenomics, this work establishes the utility of a high-throughput phenotypic analysis for further leafy crop research and improvement.

The online version contains supplementary material available at 10.1186/s12870-026-08380-6.

Figures