Pepper (Capsicum) pilot project
The collection of large-scale phenotypic data has been gaining momentum in the last couple of years. With the fast array of different screening technologies, we are capable of measuring plant phenotypes such as fruit size and colour, presence of pathogens and the effects they have on the plant. These phenotypes are classified using AI / machine learning technology or used as training sets to further develop these technologies. Also, phenotypes, including flowering and branching, are linked to biological information such as genomic data via genome-wide association studies (GWAS) and gene expression data (co-expression networks, etc). However, these (very) large phenotypical data set open new ways of approaches as well. In this project we have started to explore the possibility of the phenotypic data produced with the gantry system, on which the Phenospex PlantEye Dualscan captures as well the 3D shape as the colour of the plants on the flood-and-drain tables.
Within this pilot project we have collected 3D point clouds to extract phenotypic traits such as biomass and plant height of 60 pepper plants. We have investigated the possibility of using these 3D models as the input for the machine learning algorithms to train and predict branching events. This project provides important information for future proposals and collaborations as well as key insights into the possibilities and limitations of the NPEC facilities and, most importantly, share knowledge of the data and analytics within the Business Unit Bioscience within the Plant Science group of Wageningen University & Research (WUR).
We used the PlantEye system for phenotyping. This system delivers information such as plant height, leaf area index and 3D point clouds with RGB and near-infrared information.
The project was initiated by Plant Breeding and Applied Bioinformatics, involving, amongst others, Sander Peters, Richard Immink, Aurin Vos, Ruud de Maagd, Sara Diaz Trivino, Theo van Hintum and Willem van Dooijeweert. Data processing and analysis were performed by Christina Papastolopoulou and Sven Warris.