Read full article on the Berkeley Lab News Center.
To harness biological systems (plants and microbes) for next-generation energy production and advanced materials, researchers are looking to beneficial plant-microbe interactions. Because these are complex systems, it has proven difficult to reproducibly control exactly which microbes are present. And, subtle differences in materials, methods, or even the hands of the researchers themselves can lead to inconsistent results. This makes it difficult to replicate previous work, significantly slowing the leap from scientific discovery to practical application.
Biosciences Area researchers in collaboration with other Berkeley Lab teams are overcoming this bottleneck by addressing a multi-layered challenge: building reliable physical hardware, engineering accurate visual sensors, and developing predictive algorithms. Their solution, EcoBOT, stands out from typical plant phenotyping facilities by integrating these distinct components into a reliably automated workflow under strictly sterile conditions.
By using robotics and AI to standardize plant-microbe studies, we are building the foundational tools to accelerate science to address pressing DOE Missions and global challenges. — Trent Northen
EcoBOT takes specialized growth chambers, called EcoFABs, and integrates them with machine-learning tools that autonomously guide the discovery cycle. This system uses advanced imaging to regularly scan the entire plant—from the tips of its leaves to the bottom of its roots. By using Gaussian Process models and AI analysis tools, it can quickly analyze and model this visual data to calculate the most informative next steps. This directs the automated hardware to determine exactly how plants adapt to environmental stressors, establishing the crucial microbe-free baseline needed to eventually study plant-microbe interactions and engineer better bioenergy crops.
To illustrate how EcoBOT couples continuous measurement, adaptive modeling, and experimental redesign, the Berkeley Lab researchers used the system to observe how the model grass Brachypodium distachyon responds to environmental stressors such as nutrient deprivation and copper toxicity. In a traditional workflow, researchers might test a random spread of copper concentrations and wait weeks to measure the results. But inside EcoBOT’s compact cabinet, a robotic arm can autonomously manage over 150 individual EcoFABs simultaneously across three shelves. This robotic hardware doesn’t just automate the process; it intentionally maintains a highly controlled physical environment, providing the necessary foundation for the system-level modeling and downstream adaptive decision-making.
Historically, extracting continuous data from that many biological environments would have been a grueling, manual task prone to human error. To solve this, researchers equipped EcoBOT with a suite of Berkeley Lab-developed deep learning tools that serve as the system’s digital eyes. Below ground, a tool called RhizoNet serves as an automated root tracker. Rather than relying on inconsistent manual interpretation of root images, RhizoNet uses neural-network-based segmentation to digitally separate fragile plant roots from the noisy background of the hydroponic fluid in a standardized and reproducible way. Above ground, a computer vision tool called EcoSpec scans the plant’s shoots and analyzes complex, multi-wavelength hyperspectral images to monitor plant health. This tool has demonstrated high accuracy in high-throughput monitoring—while maintaining consistency across longitudinal measurements.
The EcoBOT becomes a true self-driving laboratory through the continuous interaction between its physical infrastructure, sensing systems, and adaptive modeling framework. The robotic hardware stabilizes the experimental environment, the imaging systems convert plant behavior into quantitative measurements, and gpCAM uses those measurements to identify where uncertainty is highest and determine which experiments should be performed next. Using Gaussian-process-based modeling, gpCAM analyzes preliminary experimental results, estimates uncertainty across the experimental landscape, and calculates the next experiments that are likely to be most informative.
“This entire platform is a great example of multi-disciplinary team science,” said Trent Northen, EGSB Deputy Division Director, who also serves as principal investigator of the m-CAFEs Science Focus Area, and co-developer of the EcoFAB and EcoBOT systems. “By using robotics and AI to standardize plant-microbe studies, we are building the foundational tools to accelerate science to address pressing DOE Missions and global challenges.”
By iteratively targeting these knowledge gaps, this autonomous approach improved the predictive accuracy of the plant biomass models by more than thirty percent. Training and processing the complex visual data for these advanced machine learning models requires massive computational power, which the team accesses using supercomputers at the National Energy Research Scientific Computing Center (NERSC).
Berkeley Lab’s culture of team science was essential to realizing this vision. Bringing the self-driving lab to life required a collaboration of plant biologists, robotics engineers, and mathematicians from the Lab’s Center for Advanced Mathematics for Energy Research Applications (CAMERA). These include BioEPIC researchers from the Biosciences Area: Trent Northen, Environmental Genomics and Systems Biology (EGSB) Division Deputy Director and co-developer of the EcoFAB and EcoBOT systems; EGSB research scientist Peter Andeer, who contributed to the design of EcoFABs and EcoBOT; EGSB staff scientist Benjamin Bowen, and EGSB research scientist Vlastimil Novak.
The development of EcoBOT was supported by several DOE Biological and Environmental Research (BER) program projects over the years. It was originally developed by the TEAMS initiative, and is now supported by m-CAFEs, the JGI, and TWINS.
