Study shows antibiotic-resistant bacteria adapt genes in WWTPs

An AI model trained on large amounts of genetic data can now predict when bacteria will become antibiotic-resistant. The new study shows that antibiotic resistance is more easily transferred between genetically similar bacteria, and that this primarily occurs in wastewater treatment plants (WWTPs) and inside the human body. This is revealed by a study from the University of Gothenburg in Sweden.

In a new study published in Nature Communications, researchers from Chalmers University of Technology, the University of Gothenburg, and the Fraunhofer-Chalmers Centre developed an AI model to analyse historical gene transfers between bacteria using information about the bacteria’s DNA, structure, and environment. The model was trained on the genetic data of nearly one million bacteria—a vast dataset compiled by the international research community over many years. One of the study’s findings is a new perspective on where resistance arises. The study reveals the environments in which resistance genes are transferred between bacteria and what makes some bacteria more prone to gene exchange than others.

Exchanging genes in WWTPs

“We see that bacteria found in humans and in wastewater treatment plants have a higher likelihood of becoming resistant through gene transfer. These are simply environments where many bacteria carrying resistance genes encounter others, often in the presence of antibiotics,” says David Lund, PhD student at the Department of Mathematical Sciences at Chalmers and the University of Gothenburg, in a press release. A fundamental reason for the rapid spread of antibiotic resistance is bacteria’s ability to exchange genes with one another, including those that make them resistant.

Combat the spread of antiobiotic resistance

“By understanding how bacterial resistance emerges, we can better combat its spread. This is crucial for protecting public health and maintaining healthcare’s ability to treat infections,” says Erik Kristiansson, professor at the Department of Mathematical Sciences at Chalmers and the University of Gothenburg in the same press release. “Harmful bacteria in humans have today have accumulated a large number of resistance genes, which is the end result of a complex evolutionary process. Many of these genes originally come from harmless bacteria that live in our bodies or the environment. Our research investigates how these genes are transferred to pathogenic bacteria, enabling us to predict how resistance might develop in future bacteria,” Kristiansson continues.

Large datasets available

The samples from wastewater treatment plants that formed the basis for the study were generally taken from three positions (beginning, middle, and end) in the aeration basins of the WWTP’s. The researchers point out that AI is at its best in complex contexts involving large datasets. “One unique aspect of our study is the very large dataset we used to train the model, which demonstrates that AI and machine learning are powerful tools for describing the complex biological processes that make bacterial infections difficult to treat,” says David Lund. Another key factor increasing the likelihood of resistance genes ‘jumping’ from one bacterium to another is genetic similarity between bacteria. When a bacterium takes up a new gene, energy is required to store the DNA and produce the protein the gene encodes, which is a cost to the bacterium.

More research for better understanding

“Most resistance genes are shared between bacteria with similar genomic structures. We believe that the cost of acquiring a new gene is then lower. We are continuing our research to more precisely understand the mechanisms governing this process,” says Erik Kristiansson. To test the model’s performance, it was evaluated against bacteria for which the researchers already knew gene transfer had occurred, but this information was withheld from the AI model—like an exam where only the researchers had the answer key. The conclusion was that in four out of five cases, the model could predict whether a gene transfer would occur. According to Erik Kristiansson, future models could be even more accurate—both through advances in the AI itself and by training it on even more data.

Practical measures to prevent gene transfer

“AI and machine learning make it possible to efficiently analyse and interpret the enormous amounts of data available today. This allows us to truly work data-driven to answer complex questions we’ve long struggled with—and to ask entirely new ones,” he says. The researchers hope that in the future, the AI model can be used in systems that quickly identify whether a new resistance gene is likely to be transferred to pathogenic bacteria, and translate that into practical measures. “For example, it could lead to better molecular diagnostics for identifying multi-resistant bacteria,  or monitoring of wastewater treatment plants and other environments where antibiotics may be present,” says Erik Kristiansson.

Antibiotic resistance rapidly increases

Antibiotic resistance is rapidly increasing as bacteria evolve the ability to withstand treatments that once killed them, making common infections harder to treat. This growing threat is fueled by overuse of antibiotics in humans and animals, and by environments like wastewater treatment plants where resistant bacteria and genes can easily spread. According to an international study from 2019 in seven European countries, the use of antibiotics mirrors concentrations of resistant bacteria in wastewater. Higher concentrations of resistant bacteria are found in wastewater treatment plants in Southern European countries.

This article was published first on the Swedish platform for water professionals cirkulation.se