The
rapid expansion of high-throughput phenotyping (HTP) platforms in agronomic
research has led to a major shift in plant science towards time-series
phenotyping that can track plant development through its life stages, providing
an opportunity to dissect the genetic basis of longitudinal traits. Plant
breeders can now phenotype large populations during the growing season and
promote the desirable genetic gain for the traits of interest in specific time points
within their breeding program. The biggest challenge is to use the various
tools in a practical way to understand the many complexities of plant growth
and development and breeding implications. This dissertation explores interdisciplinary
frameworks to assess different applications of HTP for longitudinal traits in soybean
breeding. We provide a review outlining the current analytical approaches in
quantitative genetics and genomics to adequately use high-dimensional phenomic
data. Examples, advantages, and pitfalls of each approach, and future research
directions and opportunities are explored. Among longitudinal traits in
soybean, average canopy coverage (ACC) and above-ground biomass (AGB) are
promising traits to strategic improve yield gain. Soybean ACC is highly
heritable, with a high genetic correlation to yield and can be effectively measured
by unmanned aerial systems (UAS). This study reveals that progeny rows
selection using yield given ACC (Yield|ACC) selected the most top-ranking lines
in advanced yield trials, which emphasizes the value of HTP of ACC for selection
in the early stages of soybean breeding. In addition, we developed a HTP methodology
to predict soybean AGB over several days after planting (DAP) and assessed the
quantitative genomic properties of temporal AGB using random regression models
(RRM). Our results show that AGB narrow-sense heritability estimates fluctuated
over time and the genetic correlation of AGB between DAP decreased as the days
went further apart. Considering the trait heritability, high prediction accuracies
suggest that AGB is a good indicator trait for genomic selection in soybean
breeding. Different genomic regions were found to be associated with AGB over
time with potential time-specific SNPs playing a role in the trait expression.
Similarly, candidate genes were identified with potential different patterns of
expression over time. This study presents novel genetic knowledge for longitudinal
traits in soybean and may contribute to the development of new cultivars with high
yield and optimized AGB. This is the first application of RRM for genomic
evaluation of a longitudinal trait in soybean and provides a foundation that
RRM can be an effective approach to understand the temporal genetic
architecture of a longitudinal trait in other crops.