Articles
Machine learning-driven identification of key genes regulating carotenoid metabolism and flower color in rose petals
Article number
1461_3
Pages
19 – 24
Language
English
Abstract
Carotenoid metabolism plays a key role in the yellow-to-orange coloration of rose (Rosa spp.) petals.
However, its regulatory network remains unclear, limiting progress in targeted flower color breeding.
To address this issue, we integrated transcriptomic data and carotenoid analysis data from three developmental stages of rose cultivars with different flower colors.
Using machine learning methods, including time-series clustering, PCA/t-SNE dimensionality reduction, and feature selection based on chi-square tests, we identified core gene expression modules associated with carotenoid accumulation.
K-means clustering revealed four distinct co-expression patterns, which were significantly associated with high carotenoid content phenotypes (p=0.0003). Further analysis identified candidate genes within the key modules driving flower color formation.
This study establishes a machine learning framework for gene regulatory network inference and provided functional targets for molecular breeding of novel flower colors in roses.
However, its regulatory network remains unclear, limiting progress in targeted flower color breeding.
To address this issue, we integrated transcriptomic data and carotenoid analysis data from three developmental stages of rose cultivars with different flower colors.
Using machine learning methods, including time-series clustering, PCA/t-SNE dimensionality reduction, and feature selection based on chi-square tests, we identified core gene expression modules associated with carotenoid accumulation.
K-means clustering revealed four distinct co-expression patterns, which were significantly associated with high carotenoid content phenotypes (p=0.0003). Further analysis identified candidate genes within the key modules driving flower color formation.
This study establishes a machine learning framework for gene regulatory network inference and provided functional targets for molecular breeding of novel flower colors in roses.
Authors
L.Y. Zhao, B.X. Cheng, C. Yu
Keywords
gene expression, K-means clustering, time-series analysis, flower color
Online Articles (23)