Machine learning predicts the spread of antibiotic resistance



Genes are not only inherited at birth. Bacteria have the ability to pass genes to each other or retrieve them from their environment, through a process called horizontal gene transfer, which is one of the main culprits in the spread of antibiotic resistance.

Cornell researchers used machine learning to sort organisms according to their functions and use this information to predict with near-perfect accuracy how genes are transferred between them, an approach that could potentially be used to stop the spread of resistance. antibiotics.

The team’s paper, “Functions predict horizontal gene transfer and emergence of antibiotic resistance», Published on October 22 in Science Advances. The main author is doctoral student Hao Zhou.

“Organisms can basically acquire resistance genes from other organisms. And so it would help if we knew which organisms the bacteria were exchanging with, and not only that, but we could understand what are the driving factors that involve the organisms in this transfer, ”said Ilana Brito, Assistant Professor and Mong Family 150th Anniversary Faculty Member of Biomedical Engineering at the College of Engineering, and lead author of the article. “If we could figure out who is exchanging genes with whom, maybe that would give some insight into how this actually happens and maybe even control these processes. “

Many new traits are shared through gene transfer. But scientists have not been able to determine why some bacteria engage in gene transfer while others do not.

Instead of testing individual hypotheses, Brito’s team looked at the genomes of bacteria and their various functions – which can range from DNA replication to carbohydrate metabolism – to identify signatures indicating “Who” exchanged genes and what animated these exchange networks.

Brito’s team used several machine learning models, each of them revealing different phenomena embedded in the data. This allowed them to identify multiple networks of different antibiotic resistance genes, and across strains of the same organism.

For the study, the researchers focused on organisms associated with soil, plants and oceans, but their model is also well suited for examining organisms and pathogens associated with humans, such as Acinetobacter baumannii and E. coli, and in localized environments, such as an individual’s gut microbiome.

They found that machine learning models were particularly effective when applied to antibiotic resistance genes.

“I think one of the big takeaways here is that the bacterial gene exchange network – especially for antibiotic resistance – is predictable,” Brito said. “We can figure it out by looking at the data, and we can do better if we actually look at the genome of each organism. It is not a random process.

One of the most surprising findings is that modeling predicted many possible transfers of antibiotic resistance between bacteria associated with humans and pathogens that have not yet been observed. These probable, but undetected, transfer events were almost exclusive to bacteria associated with humans in the gut microbiome or oral microbiome.

The research is emblematic of Cornell’s recently launched Center for Antimicrobial Resistance, according to Brito, who sits on the centre’s steering committee.

“One can imagine that if we can predict how these genes spread, we could either step in or choose a specific antibiotic, depending on what we see in a patient’s gut,” Brito said. “More generally, we can see where certain types of organisms are predicted to transfer with others in a certain environment. And we think there might be new antibiotic targets in the data. For example, genes that could potentially cripple these organisms in terms of their ability to persist in certain environments or to acquire these genes.

Juan Felipe Beltrán, Ph.D. ’19, contributed to the research.

The research was supported by the National Institutes of Health, the National Science Foundation, and the US Department of Agriculture.



Leave A Reply