Back to search

Cornell University

Quality trait prediction in cassava and sweet corn

Abstract

dc:description.abstract

As the world's food supply faces the dual threats of population growth and climate change, many have rallied around a call for increased productivity and plant resiliency. While this need is indisputably clear, focusing solely on quantity omits an aspect essential to true food security -- food quality. To holistically meet the dietary needs of nine billion people in the next thirty years, improved crop varieties must adhere to consumer preferences and provide adequate quantities of micronutrients. The studies described here attempt to address the challenge of developing high quality crops through genetic dissection and prediction of vitamin accumulation in fresh sweet corn kernels and the development of low-cost spectroscopy-based tools for the prediction of quality traits in cassava and beyond. Recognizing a need for open-source spectroscopy analysis software, I developed an R package, waves and integrated it into the Breedbase family of plant breeding databases as a web-based graphical interface. Easily integrated into plant breeding workflows, these tools have increased the accessibility of spectral data analysis for plant breeders. After completing tool development, the software was in the IITA cassava breeding program. Using low-cost, handheld spectrometers, spectral scans were collected from cassava roots and paired with oven-based root dry matter content phenotypes for prediction in waves. Within-trial prediction models achieved predictive abilities within the range of laboratory-grade spectrometers (R2p = 0.62 - 0.89). Given appropriate model calibration, the methods developed and tested in this study will allow for more efficient quality trait phenotyping in cassava. Finally, I explored the nutritional quality of fresh sweet corn kernels by employing a new, high-quality transcriptomic dataset for association and prediction analyses with fresh kernel carotenoid and tocochromanol traits. Though transcriptome-wide and pathway-level association analysis, significant associations were identified with six a priori candidate genes, β-carotene hydroxylase (crtRB1), lycopene epsilon cyclase (lcyE), γ-tocopherol methyltransferase (vte4), homogentisate geranylgeranyltransferase (hggt1), deoxy-xylulose synthase2 (dxs2), diphosphocytidyl methyl erythritol synthase2 (dmes2), cytidine methyl kinase1 (cmk1), and geranylgeranyl hydrogenase1 (ggh1). Transcriptomic information increased prediction accuracy over genome-wide markers for some traits, but candidate gene-targeted subsets of expression data resulted in lower prediction accuracies compared to all other models for most traits. Together, these studies represent several steps forward towards the improvement of quality trait phenotyping methods for plant breeding programs.

Degree

thesis:*
Name thesis:degree_name
Ph. D., Plant Breeding
Level thesis:degree_level
Doctor of Philosophy
Discipline thesis:degree_discipline
Plant Breeding
Grantor
Cornell University
Year dc:date.issued
2021

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Hershberger, Jenna
Chair dc:contributor.chair
  • Gore, Michael Allen
Committee members dc:contributor.committeemember
  • Setter, Tim
  • Nelson, Rebecca J.

Subjects

dc:subject × 5

Rights

dc:rights
Statement dc:rights
  • Attribution-NonCommercial-NoDerivatives 4.0 International
Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Bibid
15160200
Dc Identifier Other
ProQuest Submission ID: 12635
ProQuest Publication ID: 28649779

Chain of custody

source
Harvested from
Cornell University
Base URL
ecommons.cornell.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Hershberger, Jenna. Quality trait prediction in cassava and sweet corn. Doctor of Philosophy thesis, Cornell University, 2021. https://hdl.handle.net/1813/110560