Suddenly, a glut of data revealed hitherto unknown microbial diversity. In 2009, fewer than 1,000 bacterial genomes had been completely sequenced. By 2014, there were more than 30,000. That figure has since ballooned: At the end of 2023 there were 567,228 complete bacterial genomes, easily browsable and available for cross-reference. Today bacteria account for nearly 80% of all available genomic data.

“People just had no idea how many species there would be,” said Gralka, who now runs his own lab at VU University in Amsterdam. “You can’t tell them apart very well under the microscope.”

However, identifying individual bacterial species in a community can tell scientists only so much. Their names don’t necessarily say much about what each bug is contributing or how the community fits together.

“These communities are high-dimensional,” said Jacopo Grilli, a theoretical microbial ecologist and ex-physicist at the Abdus Salam International Center for Theoretical Physics in Trieste, Italy. “If we try to understand [them], we have to deal with the fact that there are many, many populations, many different species — whatever ‘species’ means — in these communities. All these species have all their own peculiarities, and somehow they’re coexisting.”

In 2018, a Science paper by Sanchez and his team gave microbiologists permission to simplify their thinking. Their breakthrough research showed that if you took a step back and let highly specific details, like exact species names, melt away, you could better understand the logic of a bacterial community, as if you were viewing an abstract painting from a distance.

Like Grilli, Sanchez was a physicist before turning to microbial ecology. “I decided to start working on ecology and microbial communities because I noticed that at the quantitative level, it was an area that had not been as well studied as evolution,” Sanchez said.

For the study, his lab grew wild bacteria cultured from dead leaves and soil around New Haven, Connecticut. They found that given the same set of environmental conditions — the same carbon sources, temperature, acidity and so on — any microbial community will arrive at roughly the same functional composition, no matter how it started. In his experiments, with every population, the same niches appeared and were filled over and over, though not necessarily by the same species of bacteria.

The research changed how microbiologists looked at community. When Sanchez compared communities sampled from the same environment, the names of the bacteria were always different, D’Souza said. “But if you look at the functional gene content, like who does what? That is surprisingly similar,” he said. “So it doesn’t matter who you are; what you do matters.”

The Genome’s Predictive Power

In 2018, Gralka had just arrived in Boston to work as a postdoc in Cordero’s lab at MIT. He’d started out as a biophysicist, studying the physical properties of cells, individually and in aggregates. He had decided to join Cordero’s research program because the two researchers had similar visions: to develop a quantitative, bird’s-eye understanding of microbial communities.

Cordero had a freezer stocked with Atlantic Ocean microbes, which his lab had used to make an interesting discovery about how microbial communities form around food sources, published in Current Biology in 2019. They had dropped balls of chitin — a polymer of repeating sugar molecules that makes up insect shells — into cultures of bacteria grown from the marine samples. When the scientists fished the balls back out, they looked at what communities had formed. Chitin-eating microbes were predictably clinging to the chitin — but there were also bacteria that didn’t eat chitin. Those bacteria seemed to eat the byproducts cast off by the chitin-eaters. The chitin-eaters and byproduct-eaters had formed a community.