Deciphering Cell-Type-Specific Transcriptional Regulation in Tomato Leaves Through Ensemble Machine Learning and...

Area of Science:

• Plant Molecular Biology
• Computational Biology
• Genomics

Background:

• Single-cell RNA sequencing (scRNA-seq) has revolutionized plant transcriptional landscape studies.
• Decoding cell-type-specific regulation in non-model crops like tomato (Solanum lycopersicum) presents significant challenges.

Purpose of the Study:

• To develop and apply an integrated computational pipeline for analyzing tomato leaf single-cell transcriptomes.
• To identify cell-type-specific regulatory programs and candidate transcription factors (TFs) in tomato.

Main Methods:

• Utilized high-dimensional weighted gene co-expression (hdWGCNA) and ensemble machine learning (XGBoost) on scRNA-seq data.
• Performed unsupervised clustering to identify cell subpopulations and cell-type-specific gene modules.
• Conducted in silico knockout (KO) and CellOracle simulations to predict TF functions and regulatory impacts.

Main Results:

• Identified 19 cell subpopulations across five major cell types: mesophyll, guard, trichomes, vascular, and lamina epidermis.
• hdWGCNA revealed cell-type-specific modules, linking mesophyll cells to photosynthesis and guard cells to redox homeostasis.
• Identified 33 core TFs, with four candidates (SlWRKY-78, SlWRKY-75, SlERF-57, SlGLK-49) predicted to dysregulate key pathways upon KO. Virtual deletion of SlWRKY-78 and SlWRKY-75 shifted guard cells towards mesophyll territory.

Conclusions:

• The study provides a robust computational framework for dissecting cell-type-specific regulatory networks in horticultural crops.
• Identified key TFs that play crucial roles in maintaining guard cell identity and function.
• Findings offer potential targets for genetic manipulation to improve crop traits.