Genomic Science Program
U.S. Department of Energy | Office of Science | Biological and Environmental Research Program

The Twin Ecosystems Project: A New Capability for Field and Laboratory Ecosystems Coupled by Sensor Networks and Autonomous Controls


Sheryl L. Bell1, Elle Barnes2,3, Jonathan J. Silberg4, Caroline Masiello4, Jiwoo Kim4, Li Chieh Lu4, Peter Andeer2, Peter Zwart2, Marcus Noack2, Daniela Ushizima2, James Sethian2,5, Kirsten Hofmockel1, Susannah G. Tringe2,3* (, and Trent Northen2, 3


1Pacific Northwest National Laboratory (PNNL), 2Lawrence Berkeley National Laboratory (LBNL), 3DOE Joint Genome Institute, 4Rice University, and 5University of California–Berkeley


The goal of the TWIN ecosystem project (TWINS) is to pilot self-driving lab, positron emission tomography (PET), microbial biosensors, and laboratory and field twin ecosystems to gain insights into above- and belowground plant dynamics and interactions. Autonomous experiments are used to gain novel insights into grass responses to nutrient stress, and PET is being used to collect hot spots for omics analyses. These efforts are being used to study compositional changes in root exudates and rhizosphere communities following harvest. Here the field twin defines the climate conditions for the lab twin—providing powerful environmental controls and measurements, which are essentially not possible in the field.


This project has integrated computer vision software and autonomous experimental design software (gpCAM) developed by the Center for Advanced Mathematics for Energy Research Applications (CAMERA) with an automated experimental system for performing fabricated ecosystem experiments (the EcoBOT). Three rounds of experiments have now been performed to map the nutritional landscape of the JGI flagship model grass, Brachypodium distachyon, and the hyperspectral signatures of plant combinations of nutrient stresses. The resulting model will be a valuable tool in interpreting ongoing remote multispectral image data from a field site in Prosser, Washington.  At this field site, researchers have leveraged an existing field experiment to define climate conditions for a controlled laboratory twin that replicates field conditions for plants that have been transplanted from the field into large-scale mesocosm environments (EcoPOD, the lab twin).

The laboratory twin enables detailed control and characterization of the composition and dynamics of microbes and exudates under baseline conditions and in response to perturbation. In a pilot experiment, TWINS is investigating how plant biomass harvest alters the soil microbial community structure and function in response to tall wheatgrass (Thinopyrum ponticum) exudates. Samples were collected from both the field and the EcoPOD under twinned environmental conditions in July 2022, three days prior to harvest of the plant’s aboveground biomass and again 3 days after the harvest. Bulk and rhizosphere soils were isolated and extracted for nucleic acids and polar metabolites. Amplicon sequencing of the 16S V3–V4 region was used to investigate microbial community responses to plant harvest. Root, rhizosphere, and bulk soil samples have been analyzed using liquid chromatography–tandem mass spectrometry including both reverse phase and hydrophilic liquid interaction chromatography. Untargeted metabolomic analysis has been used to compare thousands of unique chemical features to identify changes in root and rhizosphere metabolites following harvest and to compare these findings between the lab and field twins.

Engineered microorganisms could potentially be used to monitor in situ processes in fabricated ecosystems. To investigate whether information can be transmitted across a soil using a rare volatile metabolite researchers evaluated whether one member of the Model Soil Consortium-2 (MSC-2; Variovorax) can be programmed to produce a unique methyl halide signal. By expressing a methyl halide transferase in Variovorax, researchers showed that this microbe can synthesize a signal that is more than 100–fold higher than the basal level produced by other soil microbes in MSC-2 including Dyadobacter, Ensifer, Rhodococcus, and Streptomyces. The signal generated by Variovorax could be read out directly using gas chromatography or using a Methylorubrum biosensor that produces a fluorescent output. Using soil habitats ranging in size from 1 to 50 grams of soil researchers found that methyl halide cell-cell signaling could be achieved under environmentally relevant water holding conditions. Synthetic methyl halide signaling is expected to simplify fundamental studies of gene expression in hard-to-image materials containing microbiomes, and it should be useful for programming soil consortia to convert information sensed in subterranean settings into overt aboveground visual or gas signals in fabricated ecosystems.

Funding Information

The team gratefully acknowledges funding from the U.S. Department of Energy (DOE) Office of Science, Office of Biological and Environmental Research, Genomic Science Program under contract number DE-AC02-05CH11231. Work at PNNL and LBNL was performed under FWPs 78814 and FP00013570, respectively. PNNL is a multiprogram national laboratory operated by Battelle for the DOE under Contract DE-AC05-76RLO 1830.