Identifying Genes Linked to Variation in Metabolic and Whole Plant Phenotypes using Data from Genome Resequencing, Transcriptomics, and Metabolic Profiling of a Field-Grown Maize Diversity Panel

Authors

Ramesh Kanna Mathivanan, University of Nebraska-LincolnFollow

First Advisor

James C. Schnable

Committee Members

Toshiro Obata, Jinliang Yang, Souparno Ghosh

Date of this Version

7-30-2024

Document Type

Thesis

Citation

A thesis presented to the faculty of the Graduate College at the University of Nebraska in partial fulfillment of requirements for the degree of Master of Science\

Major: Agronomy

Under the supervision of James C. Schnable

Lincoln, Nebraska, August 2024

Comments

Copyright 2024, Ramesh Kanna Mathivanan. Used by permission

Abstract

Maize metabolism is highly complex and influenced by genetic variation, yet the specific genes contributing to this variation and their links to non-metabolic traits remain less understood. To address this knowledge gap, we identified genes involved in maize metabolic variation and linked them to non-metabolic traits. We utilized a quadruplicate dataset of whole genome resequencing, transcriptomic, metabolic, and whole plant phenotype data from a single common field experiment of 660 diverse maize inbred lines. Leaf samples were collected shortly before flowering and analyzed using GC-MS for 26 metabolites. A Resampling Model Inclusion Probability Genome-Wide Association Study (RMIPGWAS) of approximately 2.6 million SNPs was conducted for these metabolites, identifying 155 candidate genes, with 17 showing particularly strong signals. A parallel Transcriptome Wide Association Study (TWAS) identified 6 candidate genes. A random forest feature importance-based approach identified one overlapping gene, Cu(2+)-exporting ATPase, and other genes not found by TWAS. Three loci associated with metabolite traits were also linked to non-metabolic traits in RMIPGWAS of 41 non-metabolic traits, including whole plant phenotypes, hyperspectral, and photosynthetic traits. Key genes identified include Zm00001eb270570 (Theobromine synthase), Zm00001eb354560 (Ubiquitin carboxyl-terminal hydrolase), and Zm00001eb051410 (N-acetyl-gamma-glutamyl-phosphate reductase). Our analysis showed that each method identified unique sets of genes associated with metabolite variation, demonstrating the complementary nature of different genomic approaches. The use of machine learning techniques like RF is crucial for identifying genes from gene expression data. These findings facilitate further studies on the roles of these metabolites and genes in plant growth and development.

Advisor: James C. Schnable