Scientists at the U.S. Department of Energy’s Oak Ridge National Laboratory have developed a platform that could transform the way researchers engineer microbes for biotechnology. By combining genome shuffling, automated experiments, artificial intelligence and quantitative genetic mapping, the team can rapidly identify the tiny DNA changes that make bacteria better at producing chemicals, breaking down plant material or capturing critical minerals. The work offers a faster route to designing microbial “factories” for industrial manufacturing and could help strengthen domestic supply chains for materials now sourced through vulnerable global networks.
The platform addresses a long-standing challenge in microbial engineering. Bacteria are extraordinarily useful because they grow quickly and can manufacture enzymes, fuels, specialty chemicals and other products, but their traits are often controlled by many genes acting together. Conventional approaches typically examine what happens when an entire gene is added, deleted or switched off. That strategy can reveal major biological functions, but it may miss the subtle effects of individual DNA changes. The Oak Ridge method instead focuses on differences at the nucleotide level—the single-letter variations in DNA that can determine whether a bacterial strain performs efficiently or poorly.
To create the genetic diversity needed for this analysis, the researchers revived and modernized a technique first developed in the 1970s: protoplast fusion. Bacterial cells normally reproduce asexually, meaning that they do not produce offspring by combining genetic material from two parents in the way plants and animals do. In protoplast fusion, the protective cell walls of bacteria are removed, allowing cells from different strains to merge. Their genetic material can then recombine, producing a large population of descendants with shuffled genomes. These descendants, known as recombinants, provide the variation required to connect inherited DNA differences with measurable biological traits.
The team initially applied the approach to Bacillus subtilis, a model organism widely used in fermentation, enzyme production and agricultural biotechnology. The researchers generated genetically varied populations by crossing distinct Bacillus strains and then measured traits across many individual isolates. They also demonstrated that related genome-shuffling strategies could be used with other bacteria, including Clostridium thermocellum, which breaks down and ferments plant cellulose; Novosphingobium aromaticivorans, which processes aromatic compounds derived from lignin; and Stutzerimonas stutzeri, a versatile species associated with bioremediation, soil nutrient cycling and plant health.
This process is known as quantitative trait locus, or QTL, mapping. In QTL analysis, scientists compare the genetic variation and physical characteristics of a large population of offspring. Statistical algorithms search for DNA regions whose variants consistently correlate with a trait, such as faster growth, improved chemical production or enhanced tolerance to industrial conditions. The method allows researchers to examine many genetic differences simultaneously and narrow the search to candidate genes or regulatory sequences that may control complex characteristics. Until now, QTL mapping has been far more common in plants and other organisms with natural sexual reproduction than in bacteria.
Automation and computer vision were essential because the researchers had to phenotype—measure the biological traits of—large numbers of bacterial recombinants. A robotic system repeatedly positioned culture plates with precise control over angle, location and lighting. Consistent imaging conditions allowed high-resolution photographs to be compared reliably across the entire population. According to the researchers, automation made phenotyping roughly 10 times faster than conventional manual procedures. Artificial intelligence and computer-vision software then analyzed the images, extracted measurable traits and converted visual information into data suitable for statistical mapping.
The resulting workflow links several demanding stages that are often separated in microbial research. Scientists first create a genetically diverse population, grow and characterize the individual isolates, extract and sequence their DNA, and use statistical models to connect DNA variants with traits. Candidate genetic changes are then tested directly with CRISPR-based editing. By swapping or modifying specific sections of bacterial DNA, the team can determine whether a suspected variant actually causes the observed effect rather than merely appearing alongside it. This validation step is crucial because microbial traits may be influenced by interactions among multiple genes, environmental conditions and changes in gene regulation.
The technology could be particularly valuable for converting plant biomass into useful products. Lignin, a tough structural polymer in plant cell walls, contains energy-rich aromatic molecules but is difficult to process economically. Bacteria capable of breaking down lignin-derived compounds could be engineered to transform them into high-value chemicals, reducing waste from bioenergy and agricultural operations. Other versions of the platform could help identify microbes that absorb or concentrate critical minerals, potentially supporting cleaner recovery methods for elements needed in electronics, renewable-energy systems and advanced manufacturing.
The researchers describe the platform as a multidisciplinary achievement involving robotics, microbial genetics, genome sequencing, image analysis, artificial intelligence and statistical biology. The work was led by Josh Michener and Dan Jacobson at Oak Ridge National Laboratory, with support from the laboratory’s research and development program, the DOE Center for Bioenergy Innovation and the Secure Ecosystem Engineering and Design Science Focus Area. Genome sequencing was performed through the Joint Genome Institute, a DOE Office of Science user facility. The microbial QTL mapping technology is available for licensing, and researchers are already applying it to industrial microbes, plant-associated bacteria and airborne microbial communities. The study, published in Nature Communications, signals a shift toward faster, more precise and increasingly automated control over the microbial systems that could manufacture the next generation of materials.
Article Title: Genome shuffling enables quantitative trait locus mapping in Bacillus subtilis
Web References: Nature Communications article; ORNL technology licensing page; ORNL background on protoplast fusion
References: Nature Communications, DOI: 10.1038/s41467-026-72929-0
Image Credits: Andy Sproles/ORNL, U.S. Department of Energy
Keywords: Biotechnology, synthetic biology, genetic engineering, microbial engineering, quantitative trait locus mapping, genome shuffling, artificial intelligence, CRISPR, bioenergy, critical minerals, Bacillus subtilis
