SAPOR Seminar | Speaker: Maria Xose Rodriguez-Alvarez, Universidade de Vigo
Spatial variation is common in agricultural experiments, where phenotypic traits are often influenced by micro-environmental factors, such as soil heterogeneity. To address this, researchers use both experimental designs and spatial models to effectively separate genetic and non-genetic effects. Recently, high-throughput phenotyping (HTP) platforms have emerged as powerful tools for the non-destructive acquisition of phenotypic data with high spatial and temporal resolution, providing valuable insights into plant growth, development, and genotypic performance. However, analysing HTP data introduces a new challenge: integrating both spatial and temporal dimensions.
In this talk, we will discuss the use of spatial and spatio-temporal P-splines combined with mixed models for analysing phenotypic data. We will begin by considering cases where the phenotype of interest is measured only once, typically at the end of the experiment, and show how two-dimensional P-splines can be used to correct for spatial variation, enabling accurate estimation of genotypic effects. We will then extend this framework to the analysis of HTP data, where phenotypic traits are repeatedly measured over time. In this setting, three-dimensional P-splines are used to account for spatio-temporal variation, while the temporal evolution of genetic effects is modelled through genotype-specific P-spline curves. Examples from plant breeding and HTP experiments will be presented to illustrate these models.