Degree Type

Dissertation

Date of Award

2019

Degree Name

Doctor of Philosophy

Department

Agronomy

Major

Plant Breeding

First Advisor

Asheesh K. Singh

Abstract

Plant scientists are beginning to harness the capabilities of high dimensional ‘omic tools (e.g., genomic, phenomic) to usher in the era of digital agriculture to allow the usage of predictive analytics. While genomic tools have been developed to exploit high-density genetic markers for breeding decision making, a gap persists in the availability of phenomic-assisted breeding methodologies. Here we develop frameworks malleable to crop species and breeding objective to leverage complex high-dimension phenomic data using machine learning (ML) and optimization techniques for the development of data driven solutions designed to empower plant scientists to; develop prescriptive breeding solutions, improve the operation efficiency of breeding programs, and to expand the capacity of current phenotyping efforts through the use of a fine-tuned package of sensors assembled for a specific breeding objective. In this consortium of work, we show that phenomic predictors can be deployed for ML assisted prescriptive-breeding techniques for precision product placement and in turn these same phenomic predictors can be used for efficient cultivar testing (e.g., seed yield) to optimize breeding program operational efficiencies. Furthermore, phenomic sensors provided a wealth of data making this work ripe for genomic studies revealing the underlying genomic regions controlling yield predicting phenomic traits and rapid scanning of genotyped germplasm using genomic prediction. This work will allow breeders to continually optimize their breeding programs to begin fusing widely available genomic data with the upcoming capabilities of high throughput phenotyping techniques to streamline cultivar development pipelines.

Copyright Owner

Kyle Parmley

Language

en

File Format

application/pdf

File Size

128 pages

Share

COinS