Gene expression, transcription factor binding and histone modification predict leaf adaxial–abaxial polarity related genes

Abstract

Leaf adaxial–abaxial (ad–abaxial) polarity is crucial for leaf morphology and function, but the genetic machinery governing this process remains unclear. To uncover critical genes involved in leaf ad–abaxial patterning, we applied a combination of in silico prediction using machine learning (ML) and experimental analysis. A Random Forest model was trained using genes known to influence ad–abaxial polarity as ground truth. Gene expression data from various tissues and conditions as well as promoter regulation data derived from transcription factor chromatin immunoprecipitation sequencing (ChIP-seq) was used as input, enabling the prediction of novel ad–abaxial polarity-related genes and additional transcription factors. Parallel to this, available and newly-obtained transcriptome data enabled us to identify genes differentially expressed across leaf ad–abaxial sides. Based on these analyses, we obtained a set of 111 novel genes which are involved in leaf ad–abaxial specialization. To explore implications for vegetable crop breeding, we examined the conservation of expression patterns between Arabidopsis and Brassica rapa using single-cell transcriptomics. The results demonstrated the utility of our computational approach for predicting candidate genes in crop species. Our findings expand the understanding of the genetic networks governing leaf ad–abaxial differentiation in agriculturally important vegetables, enhancing comprehension of natural variation impacting leaf morphology and development, with demonstrable breeding applications.

Keywords

• Machine learning
• Leaf polarity
• Arabidopsis thaliana
• Brassica rapa
• Transcription factor