A custom pipeline for building computational models of plant tissue

Abstract

Stalk lodging in the monocot Zea mays is an important agricultural issue that requires the development of a genome-to-phenome framework, mechanistically linking intermediate and high-level phenotypes. As part of that effort, tools are needed to enable better mechanistic understanding of the microstructure in herbaceous plants. A method was therefore developed to create finite element models using CT scan data for Zea mays. This method represents a pipeline for processing the image stacks and developing the finite element models. 2-dimensional finite element models, 3-dimensional watertight models, and 3-dimensional voxel-based finite element models were developed. The finite element models contain both the cell and cell wall structures that can be tested in silico for phenotypes such as structural stiffness and predicted tissue strength. This approach was shown to be successful, and a number of example analyses were presented to demonstrate its usefulness and versatility. This pipeline is important for two reasons: (1) it helps inform which microstructure phenotypes should be investigated to breed for more lodging-resistant stalks, and (2) represents an essential step in the development of a mechanistic hierarchical framework for the genome-to-phenome modeling of herbaceous plant stalk lodging.