The ocean is full of invisible workers. Trillions of microbes quietly break down carbon-containing organic matter, which helps to regulate Earth's climate. But scientists have long struggled to understand how different microbes contribute to the process.
Now, researchers at the USC Dornsife College of Letters, Arts and Sciences and collaborators have developed a new way to make sense of that hidden workforce. Their study, published recently in Science Advances, identifies a small set of "metabolic niches" — or functional roles — that help explain how marine microbes grow, compete for resources and recycle carbon around the globe.
"These microbes are incredibly diverse, but we found that their behavior can be grouped into a manageable number of strategies," said study lead scientist Naomi Levine , professor of biological sciences, quantitative and computational biology, and Earth sciences at USC Dornsife. "That gives us a much clearer way to connect microbial life to the carbon cycle."
USC Dornsife scientists map the ocean's invisible workforce
Marine microbes play a central role in Earth's climate. Some of these single-cell organisms use photosynthesis to turn carbon dioxide into organic molecules such as sugars, while others — including the microbes Levine's team studied — consume those molecules as food, releasing much of the carbon back into the ocean as carbon dioxide. This cyclic process helps shape how much carbon the ocean stores versus sends back into the atmosphere.
But scientists have struggled to predict how these processes work because microbial communities are so complex. Thousands of species can coexist in a single bucket of ocean water.
"The big challenge has been figuring out how to simplify that complexity without losing what really matters," said Levine, who holds the Gabilan Distinguished Professorship in Science and Engineering.
To tackle this problem, the research team analyzed genetic data from thousands of marine microbes collected around the world. Using that information, they built computer models that simulate how each organism uses different types of food — sugars, amino acids or organic acids — to grow. The team then simulated how each microbe responded when certain nutrients were limited. This revealed which resources each organism depends on most.
From these patterns, the researchers used a machine learning approach to group microbes into eight broad clusters, each representing a different metabolic strategy to obtain and use nutrients.
Some clusters included fast-growing "generalists" that can use a wide range of food sources. Others consisted of slower growing "specialists" that rely on specific types of nutrients.
"It's a little like categorizing people by how they eat," Levine said. "Some will eat almost anything, while others depend on very particular diets. Those differences shape how they live and where they thrive."
The eight metabolic groups help explain how microbial communities vary across the ocean.
For example, generalists were more common in nutrient-rich environments like coastal waters, especially where rivers meet the sea. In contrast, slower-growing specialists were more prevalent in the open ocean, where nutrients are scarce.
These patterns suggest that microbial communities are structured by trade-offs. Organisms that grow quickly tend to be flexible in what they eat, while those that grow slowly are often more specialized.
Why microbial groupings matter for climate
The findings could improve how scientists model the ocean's role in the global carbon cycle.
Current climate models often struggle to represent microbial activity because of its complexity. By reducing microbial diversity into a small number of functional groups, the new framework makes it easier to include these processes in large-scale models.
That could lead to better predictions of how the ocean will respond to climate change — including how much carbon it will store in the future.
"If we want to understand climate, we have to understand the microbes," Levine said. "They're the engines driving carbon cycling in the ocean."
The study builds on earlier research, including a 2025 study done in collaboration with USC Dornsife Professor of Biological Sciences Jed Fuhrman 's group and led by Emily Zakem, a former postdoc in Levine's lab who is now at Carnegie Science.
That study used ecological models to describe how microbial communities vary across the ocean and showed that broad categories of microbes — such as fast-growing "copiotrophs" and slower "oligotrophs" — can explain large-scale patterns in carbon cycling.
The new study adds more nuance to that picture. Instead of grouping microbes based mainly on ecological traits, it identifies specific metabolic strategies based on what organisms can actually consume and how they respond to resource limits.
Together, the two studies suggest a path forward: combining ecological models with detailed metabolic information to better understand how microbial communities shape the carbon cycle.
USC Dornsife study simplifies a vast microbial world
The researchers note that their framework does not capture all microbial diversity. In particular, some groups of ocean microbes are still poorly represented because scientists lack high-quality genetic data for them.
In addition, the models rely on predictions about how microbes use nutrients, which may not fully reflect real-world behavior.
Future studies, including more laboratory experiments and improved genomic data, could refine the model and expand it to include additional microbial groups.
Still, despite these limitations, the research offers a promising step toward understanding one of the most complex systems on Earth.
By identifying a small number of metabolic strategies among these tiny but highly influential organisms, the researchers provide a new way to connect microscopic processes to global climate dynamics.
"We're trying to take something incredibly complicated and find the underlying patterns," Levine said. "Once you see those patterns, it becomes much easier to understand how the whole system works."
About the study
In addition to Levine, study authors include Ryan Reynolds, Anna Weiss, Chase James, Conner Kojima and J. Cameron Thrash from USC Dornsife, and Jackie Weissman from Stony Brook University and The City College of New York.
The study was funded by Simons Foundation grants 542389, LS-SIAME-00001961 and LS-SIAME-00001997, and National Science Foundation grants EF-2125191 and OCE-1945279.
