Plants + HPC

Many fields of research are seeing changes in how that research is conducted thanks to advances in technology. Plant science is one of those, and NCSAis playing a role in two large projects.

Historically, successful trait selection in plant breeding has involved manual measurement of individual plants. This requirement limits the number of plantsthat can be evaluated, and the scope of properties that can be measured. A new grant from the Department of Energy to researchers at the Donald DanforthPlant Science Center and multiple partner institutions, including the University of Illinois and NCSA, will fund the development of a system to automate the measurement of plants using cameras and other sensors mounted on drones, tractors, and robots, and analysis of the resulting large data sets to facilitate thedevelopment of high-yielding strains of sorghum, a key bioenergy crop.

The $8 million grant was one of several awarded by the DOE Advanced Research Projects Agency-Energy (ARPA-E) Transportation Energy Resourcesfrom Renewable Agriculture (TERRA) program. Todd Mockler, the Geraldine and Robert Virgil Distinguished Investigator at the Danforth Center, is the principal investigator. Of the total grant, $1.8 million will go to NCSA to establish a supercomputing pipeline for a reference sensing platform. Plant biologist David LeBauer will act as principal investigator for this component of the project.

LeBauer, an NCSA Fellow and Carl R. Woese Institute for Genomic Biology affiliate, will work with groups at NCSA to establish a reference data set andcomputing environment that will support all of the researchers funded by the TERRA program.

Researchers will develop a cutting-edge, automated system to collect, analyze, and share data on multiple characteristics of plants growing in the field via sensors in the air, on the ground, and mounted on tractors, and link these observations to genomic data collected from individual plants. Data collected will be used as a reference upon which new sensors, sensor platforms, and data analysis pipelines can be developed. By the project’s end, groups expect to develop a system that can survey every plant in a 50 hectare area (almost 100 football fields) each day.

Ultimately, these tools will be used to develop plants strains that are able to tolerate stresses such as drought, temperature, and disease.

Researchers at NCSA will also make the computing solutions developed for the project publicly available, allowing their future use in a variety of Big Dataresearch applications.

Other partner institutions for the grant include Clemson University, the HudsonAlpha Institute for Biotechnology, Kansas State University, Texas A&MUniversity, the University of Arizona, and Washington University in St. Louis, with key collaborators at the U.S. Arid Land Agricultural Research Center of theUSDA-Agricultural Research Service.

Plants in silico

Understanding, accurately predicting, and responding to the ways in which plants, particularly food crops, react to climate change is a critical challenge. It’s also one that encompasses multiple scales—from molecular processes to ecosystems.

NCSA is part of an interdisciplinary, international team, led by University of Illinois Crop Sciences and Plant Biology professor Steven Long, that is tackling this challenge through a project called “Plants in silico: A Multiscale Modeling Platform to Predict Crop Response to Climate Change.” The project recently received $350,000 in seed funding from the Illinois Institute for Sustainability, Energy, and Environment (iSEE).

“The rate of change of our atmosphere and climate pose a significant threat to food security by imposing environmental challenges that present-day systems are not adapted to handle,” Long wrote in the project proposal. “Designing more sustainable crops to increase productivity depends critically on complex interactions between genetics, environment, and ecosystem.

“We propose the creation of a modeling platform and a framework that would allow for the implications of a discovery at one level to be examined at the whole plant or even crop or natural ecosystem levels.” Plants in silico co-PI Amy Marshall-Colon, a U of I assistant professor of Plant Biology who specializes in plant systems biology and systems genetics, explains that models that can communicate across different biological scales have the potential to provide more accurate simulations of plant response to the environment than any single model could alone.

“What is even more exciting though, would be the ability to predict how plant interactions with their surrounding environment change in response to aperturbation directed at the plant, either genetic or environmental,” she says. “For example, we know that plant-plant and plant-microbe interactions have ahuge influence on plant development and productivity. Combining these ecosystem level models with plant cellular models has never been done before, sothere is great potential to learn just how important it is to include plant-environment interactions into crop models to improve their prediction accuracy.”

Combining models across these multiple levels is very complicated. According to Marshall-Colon, the Plants in silico team will use pioneering multiscale modeling work in the microbe and mammalian research communities as an example and will even take advantage of some of their existing tools.

“We are also fortunate that we have been able to assemble such a capable team here at the University of Illinois to take on such a venture and to build off of existing tools or create new tools when needed,” she added. The team includes NCSA director H. Edward Seidel, who has expertise in high-performance computing; Chemical and Biomolecular Engineering assistant professor Diwakar Shukla, who works in molecular modeling and simulations; Plant Biology assistant professor James O’Dwyer, who specializes in mathematical ecology; and Plant Biology professor Donald Ort and Xinguang Zhu, group leader at the Institute of Computational Biology in the Chinese Academy of Science, who are photosynthesis experts. The team also includes Mike Freemon of the National Data Service; Donna Cox, director of NCSA’s Advanced Visualization Laboratory and professor in the School of Art and Design, and Computer Science professor John Hart, who will work with researchers to develop high-resolution visualizations of integrated data.

Marshall-Colon called NCSA “the ideal collaborator to host Plants in silico.” The center will provide state-of-the-art computational and storage resources and expertise, including the National Data Service and Brown Dog technologies for data synthesis and large-scale data sharing, storage, and management. NCSA’s Advanced Visualization Laboratory is “the ideal partner to enable animated simulations of plant growth and complex ecosystem interactions,” she added.

The Plants in silico project holds the potential for tremendous impact on agriculture and food production.

“There is an immediate need today to design crops that are able to not only survive, but also produce higher yield under the inhospitable climate predictedfor 2050,” Marshall-Colon says. “Crop model predictions made possible through the Plants in silico platform will better inform researchers and direct their engineering efforts. This fine-tuning of research strategies has the potential to accelerate breeding programs and the development of higher yielding crops to feed our ever increasing and hungry population.”

Disclaimer: Due to changes in website systems, we've adjusted archived content to fit the present-day site and the articles will not appear in their original published format. Formatting, header information, photographs and other illustrations are not available in archived articles.