Computational Models

‘What’s your favorite isP article and why?’

The editors of in silico Plants are delighted to share with you an Editor's Choice selection of key papers that highlight some of the best current plant modelling research published in the journal.
Prof Steve Long

Editor in chief Dr. Stephen Long (University of Illinois and Lancaster University) has developed a range of mechanistic models at different scales relating environmental effect on photosynthesis to plant productivity. He chose the Editorial “The roles of credibility and transdisciplinarity in modelling to support future crop improvement.”

He writes:

I chose this paper because it highlights the need for dynamic plant/crop growth and development models (P/CGM) to deal with future challenges of food supply and climate change. These models predict trajectories of plant and crop attributes through the growth cycle for ExGxM scenarios. This work is exemplary of the goals of the in silico Plants journal.

The author identifies three issues that are critical to the future practice of plant modelling to support crop improvement:

  • Model Credibility – Predict phenotypic outcomes with high-quality data, generate known qualitative responses, predict emergent phenotypes and interactions associated with key traits.
  • Modeller Credibility – Biological realism grounded in solid experimental and theoretical evidence, appropriate parsimony for the application, structured to allow parameter variation, transparency, repeatability, and accessibility.
  • Transdisciplinarity – Urgent need to capture the nexus between molecular and ecophysiological understanding, a crop systems biology with an ability to scale from phenotype down and molecular up, requiring a new level of communication with non-modelers and recognition that experimenters are not simply data gatherers, but domain experts.

Going forward, our attention to these points and this key paper will be critical to move crop modeling from the periphery to the central position that it needs to occupy in the future of plant sciences and its challenges.


Prof Przemyslaw Prusinkiewicz

Professor Przemyslaw Prusinkiewicz (University of Calgary) is a pioneer in the computational modeling, simulation, and visualization of plant development.  He selected “Positional variation rather than salt stress dominates changes in 3D leaf shape patterns in cucumber canopies.”

Plots of leaf shape.

He writes:

I found this paper interesting, because I study leaves and their variability myself, and cucumber leaves have not been analyzed as widely as those of model plants Arabidopsis, Caradamine, tomato, or grapewine.  Moreover, the authors consider environmental stresses as a factor potentially affecting leaf shapes, whereas in most papers the impact of environment is ignored.

By using detailed quantification of the diversity of (cucumber) leaf forms, the authors found that leaf shape depends predominantly on leaf position in the plant. When modeled, leaf shape variation played only a minor role in light distribution in the plant canopy. The leaf (sub)model can be used in various computational plant models and is available from the authors upon request.


Prof Xin-Guang Zhu

Multi-scale systems model developer, Professor Xin-Guang Zhu (Chinese Academy of Sciences) chose “yggdrasil: a Python package for integrating computational models across languages and scales.”

A diagram of how yggdrasil works.

I chose this article because though there are many models developed by the research community; however, they cannot be easily used. I like the idea of creating a tool, which can help integrate the different computational models to describe the biological processes governing plant growth and development, covering scales from atomistic to global will not be lost.

yggdrasil has three characteristics that are important for wide use:

  • Easy to use. Require as little modification to the model source code as possible and only in the language of the model itself.
  • Efficient. Allow models to run in parallel with asynchronous communication that does not block model execution when a message is sent, but has yet to be received.
  • Flexible. Provide the same interface to the user, regardless of the communication mechanism being used or the platform the model is being executed on.

The package is available in Python, C, C++, and Matlab, and can be used on Linux, Mac OS, and Windows operating systems. yggdrasil is user-friendly and can be accessed at https://github.com/cropsinsilico/yggdrasil

>