How AI could tackle antibiotic resistance in poultry

A combination of gene sequencing and artificial intelligence (AI) could identify how antibiotic resistant bacteria is transmitted between humans, poultry and their environment.

Doughman Headshot3 Headshot
kgtoh | bigstockphoto
kgtoh | bigstockphoto

A combination of gene sequencing and artificial intelligence (AI) could identify how antibiotic resistant bacteria is transmitted between humans, poultry and their environment.

“We are interested in how antimicrobial resistance emerges and spreads in this environment,” said Tania Dottorini from the School of Veterinary Medicine and Science at the University of Nottingham.

“If we can gather enough information to understand these hot spots, we could use sophisticated computational tools like machine learning to predict where they are and prevent them from appearing again.”

Antibiotic use in poultry and other livestock is under scrutiny due to growing antibiotic resistance and consumer concerns. Resistance, when bacteria develop the ability to defeat the drugs designed to kill them, can devastate poultry flocks and affect the livelihood of farmers. 

In addition, a growing number of poultry producers are pledging to raise birds antibiotic-free or with no antibiotics ever in response to these concerns.

Isolating E. coli

For the study, published in PLOS Computational BiologyDottorini and her team of researchers collected 154 samples from live poultry, necropsies, workers and their households and the poultry houses at a large scale commercial poultry house in China.

From these samples, the researchers isolated the bacterium Escherichia coli or E. coli

E. coli is found in the intestines of most animals, including poultry and humans. The bacteria is harmless in the gut, but it can become pathogenic and result in a deadly infection if it crosses the bloodstream. The E. coli genome can also carry resistance genes against certain drugs, making the bacteria antibiotic resistant and more difficult to treat.

Identifying the bacterial genes associated with antibiotic resistance

The researchers used a computational approach that included machine learning, a form of AI, whole genome sequencing, gene sharing networks and mobile genetic elements to characterize the different types of bacteria found in the samples.

“Antimicrobial resistance is not a static phenomenon, rather it is a dynamic phenomenon. For example, one gene that has been defined as causing resistance in bacteria might become inactive, acquire a mutation or, even more worryingly, other genes that we don’t know may become the cause of resistance,” Dottorini explained.

“With artificial intelligence and machine learning, we look at the whole repertoire of genes known and unknown and monitor them to see if they change, how they change and how they are associated with resistance.”

The results showed that antimicrobial genes were present in both the pathogenic and non-pathogenic forms of bacteria present. This discovery revealed an entire network of genes associated with antimicrobial resistance (AMR), including several genes whose link to AMR was previously unknown.

“Our aim is to develop early warning solutions where everything that emerges is stores completely in the cloud and machine learning gives you potential hot spots,” she added.

The research was done in collaboration with Junshi Chen, Fengqin Li and Zixin Peng from China National Center for Food Safety Risk Assessment (CFSA).

Page 1 of 39
Next Page