To meet the growing demand for faster scientific discovery that strengthens the bioeconomy, plant scientists worked with manufacturing systems engineers at the Department of Energy's Oak Ridge National Laboratory to develop robotics and computer vision to accelerate the development of new stress-tolerant plants. The research benefits a range of applications from better bioenergy and food crops to new materials, chemicals, and for recovery of critical minerals.
The project, Sensors, Machine vision, Automation and Robotics for Transforming Plants (SMART Plant 1.0), successfully demonstrated an early stage plant transformation system that uses robotics to replace a laborious, manual task in plant science SMART Plant 1.0 identified and robotically collected tissue from live plants, treated it to transform and trigger growth, and placed it on a special nutrient blend so that it develops a cell mass that can form a new plant.
SMART Plant 1.0 aims to make the process faster, more consistent and reliable by automating this task at the heart of predictive plant biology. The system, detailed in an IEEE conference paper , could increase plant transformation efficiency by a hundredfold.
The advantage of SMART Plant 1.0 isn't just about increasing throughput, said project lead Udaya Kalluri of ORNL's Biosciences Division. "It is also about collecting massive amounts of data that can be processed with artificial intelligence to help us validate gene-to-trait linkages faster and optimize plant transformation rates. There's a huge appetite for quality data that can train AI models for plant science, and our laboratory automation work specifically supports data generation."
SMART Plant 1.0 integrates robotics, advanced computer vision-guided tissue identification and custom-developed tools for excising uniform, delicate, live plant samples with minimal damage, said project collaborator Alex Walters of ORNL's Manufacturing Sciences Division, who led the engineering systems work for the platform.
Automation eliminates inconsistencies and biases in tissue selection and handling, improving reproducibility, and ultimately enabling on-demand, autonomous plant transformation workflows.
"The plant transformation process is typically a technician working with a hole punch or scissors to gather tissue, then placing it in several growing media over successive weeks and manually recording data by hand - resulting in a slow, labor-intensive process," Walters said. "We introduced robotic excision and manipulation of tissue samples using a vision-guided system. The result was the first of its kind automation workflow for plant bioscience and heralds a future where rapid, reliable and high-throughput plant genetic engineering becomes routine and scalable."
The project also supports the scale of data generation needed to advance predictive AI models that can further expedite gene function studies for faster transformation, pushing forward agricultural innovations to bolster the U.S. bioeconomy.
Next-gen SMART Plant system expedites plant transformation pipeline
With a successful 1.0 demonstration, the ORNL team is working on a next-generation system, SMART Plant 2.0, an automated plant transformation laboratory for the high-throughput production of transgenic plants and associated AI-quality data.
SMART Plant 2.0 will collect molecular and plant trait data to enable autonomous experiments with continuously refined protocols and to enable cross-scale AI models, Kalluri said. The 2.0 project aims at enabling agentic AI systems, which automatically adjust experiments based on experiential findings, to turbocharge improvements across the entire pipeline of plant transformation and performance prediction.
The next-generation SMART Plant 2.0 system aligns with another of ORNL's automated plant science facilities, the Advanced Plant Phenotyping Laboratory, or APPL , where plants robotically rotate through an advanced array of imaging modalities that gather huge amounts of data on plant growth and function for AI-driven analysis.
ORNL's advanced new plant science facilities will integrate into a comprehensive network of plant biotechnology capabilities housed at DOE's national laboratories. The nationwide platform will be aimed at accelerating scientific discoveries for plants that tolerate challenging growing conditions and are well suited for the production of biofuels, chemicals and materials, and as a resource for critical minerals recovery and food security.
"The SMART Plant projects represent a critical leap forward to meet the growing global demand for faster genetic crop improvements by optimizing transformation steps," Kalluri said. "The project also supports the scale of data generation needed to advance predictive AI models that can further expedite gene function studies for faster transformation, pushing forward agricultural innovations to bolster the U.S. bioeconomy."
The SMART Plant projects are supported by ORNL's Laboratory-Directed Research and Development (LDRD) program. Technology related to SMART Plant is available for licensing
