DeBolt, Seth

Research Overview

The goal of this project is to understand how maize stalk and internodal morphology, composition, and material properties function to determine stalk strength and ultimately determine stalk lodging resistance.  For this purpose, over the last two years, our group has collected multilevel phenotypic data representing over 40,000 individual stalks from a maize inbred diversity panel containing 420 genotypes.  As data collection concludes, our group seeks to combine our high-resolution phenotypic dataset with pre-existing genotypic datasets in order to build models capable of identifying the most physically important and genetically tractable determinants of stalk lodging resistance.  The application of this research will allow us to better understand how best to breed for stalk lodging resistance in not just maize, but other grain crops such as rice, wheat, and sorghum – each of which suffer from billions of dollars’ worth of lost revenue due to the poorly understood causal elements of stalk lodging.

Personnel:

Norbert Bokros, PhD Candidate, Added on LCC, 01/25/2022

Virginia Verges, Postdoc, Added on LCC, 01/25/2022

Software:

R, Python, TensorFlow, PyTorch, Keras, TASSEL

R Libraries:

h2o, reshape2, tidyverse, data.table, DALEX, prospectr, caret, caretEnsemble, doParallel

Computational Methods:

Feature engineering/extraction, machine learning enabled modelling, GWAS

Grants:

Publications: