window.dataLayer = window.dataLayer || []; function gtag(){dataLayer.push(arguments);} gtag('js', new Date()); gtag('config', 'UA-29122137-2');

Machine learning algorithm improves antibiotic prescribing

Posted: 21st March 2022

Using genomic sequencing and machine learning, scientists have developed an antibiotic prescribing system that cuts resistance by half. They tested the new platform by providing personalised prescriptions for urinary tract infections (UTIs) and wound infections, improving outcomes and reducing resistance. It demonstrates the power of technology and the importance of genetics in antibiotic prescribing.

The over-prescription of antibiotics has led to a rise in antibiotic-resistant forms of bacteria that poses a risk to our health. As a result, clinicians are encouraged to prescribe antibiotics only when they’re essential and assessed to have a high chance of success against the bacteria – and they’re being supported by technology. In a new paper published in the journal Science, researchers describe how they have developed a unique insight powering an algorithm to develop personalised antibiotic prescribing.

“We wanted to understand how antibiotic resistance emerges during treatment and find ways to better tailor antibiotic treatment for each patient,” says co-author Professor Roy Kishony. They discovered that antibiotic resistance wasn’t caused by random mutations but reinfection by existing bacteria from a patient’s microbiome.

This insight that antibiotic susceptibility of past infection can be used to predict their risk of reinfection enabled researchers to go further. “Using this data, together with the patient’s demographics like age and gender, allowed us to develop the algorithm,” said co-author Dr Mathew Stracy.

The team combined microbiome profiles of over 200,000 patients with whole-genome sequencing of over 1000 bacterial samples and machine-learning analysis of 140,349 urinary tract infections and 7,365 wound infections. The team utilised advanced machine learning technologies to analyse the data and identify crucial links that humans would struggle to spot.

The new algorithm proves to be highly effective at predicting antibacterial resistance. This information can help “minimise” reinfection, say the researchers. Across a population, it offers a fascinating “means to reduce the emergence and spread of resistant pathogens,” the authors say.

The algorithm has proved its value on paper, with the next step clinical adoption, said co-author Dr Tal Patalon. “I hope to see the algorithm applied at the point of care, providing doctors with better tools to personalise antibiotic treatments to improve treatment and minimise the spread of resistance.”

At Biocentaur, we’re at the cutting-edge of personalised medical treatments. Our SNPs test uses our advanced DNA-analysis capabilities to understand how a person will respond to treatment with a range of common medications, including antibiotics.

The test uses a technique called quantitative polymerase chain reaction (qPCR) to identify genetic variations called single nucleotide polymorphisms (SNPs). These mutations provide clinicians with a vital insight into how successful antibiotics and other drugs will be. It can also identify any potentially dangerous interactions, providing peace of mind and protection.

Learn more about our entire range of tests, including how to book, here.

You can read the full paper, Minimizing treatment-induced emergence of antibiotic resistance in bacterial infections, here.

Shopping cart0
There are no products in the cart!
Continue shopping