Credit: CC0 Public Domain
Genes can be inherited from birth, but not just through the genes themselves. Bacteria can pass genes on to one another or pick up genes from their environment. Horizonal gene transfer is a key factor in antibiotic resistance spreading.
Cornell researchers used machine-learning to classify organisms according to their functions. They used this information to predict how genes will be transferred between them with near-perfect accuracy. This could help stop the spread antibiotic resistance.
The paper by the team, "Functions Predict Horizontal Gen Transfer and the Emergence of Antibiotic Resistance", was published in Science Advances on Oct. 22. Hao Zhou, a doctoral student, is the lead author.
"Organisations can basically acquire resistance genes from another organisms. It would be helpful to know which organisms bacteria are exchanging with. Not only that, but it would also help us identify the driving factors that lead to this transfer," stated Ilana Brito (assistant professor, Mong Family Sesquicentennial Faculty Fellow, Biomedical Engineering, College of Engineering and senior author of the paper. If we could identify who is exchanging genes, it might give us insight into the process and allow us to control it.
Gene transfer is a way to share many novel traits. Scientists have not been able to identify why certain bacteria engage in gene transfers and others don't.
Brito's team analyzed bacteria genomes and the functions they perform, which can include DNA replication and metabolizing carbohydrates. This allowed them to identify signatures that indicated "who" was swapping genes and what drove these exchange networks.
Brito's team used multiple machine-learning models to discover different phenomena in the data. This allowed them to identify multiple networks that contained different antibiotic resistance genes across different strains.
The study was focused on organisms that are associated with soil, plants, oceans, but the model can also be used to examine pathogens and organisms associated humans, such Acinetobacter baumannii or E. coli. It can also be used to analyze localized environments such as the microbiome of an individual.
The machine-learning models proved particularly useful when they were applied to antibiotic resistance genes.
Brito stated, "I believe one of the major takeaways from this is that the network for bacterial gene exchangespecifically for antibiotic resistanceis predictable." It is possible to understand it by looking at data. However, we can make better decisions if we look at each organism's DNA. It is not random.
One of the most unexpected findings was the prediction that there would be many potential antibiotic resistance transfer between human-associated bacterias and pathogens. This has yet to be observed. These possible, but not yet detected, transfer events were almost exclusively for human-associated bacteria in either the oral or gut microbiomes.
Brito, who is a member of the steering committee, said that this research is emblematic for Cornell's newly launched Center for Antimicrobial Resistance.
Brito stated, "One can think that if we can predict the spread of these genes, we might be in a position to intervene or select an antibiotic depending on what we see inside a patient's stomach." "We may be able to see the areas where certain organisms will transfer with others in a particular environment. We believe there may be new antibiotic targets in these data. These genes could potentially be fatal to these organisms in terms of their ability or inability to persist in certain environments.
Juan Felipe Beltrn (Ph.D.'19) contributed to this research.
Continue reading World's first study reveals that some microorganisms are capable of bending the rules for evolution
Information: Hao Zhou and colleagues, Functions predict horizontal genetic transfer and the emergence antibiotic resistance, Science Advances (2021). www.science.org/doi/10.1126/sciadv.abj5056 Journal information: Science Advances Hao Zhou et al, Functions predict horizontal gene transfer and the emergence of antibiotic resistance,(2021). DOI: 10.1126/sciadv.abj5056
