![Jessica James, second year PhD student, analyses the behaviour of bacteria using photomicrography and AI](/https/www.telegraph.co.uk/content/dam/global-health/2024/07/12/TELEMMGLPICT000385337986_17207962320720_trans_NvBQzQNjv4BqQTAoG9Ka6ECxCVPT5zp0r7ETk8P9H3T_ZmHI76ctXA8.jpeg?imwidth=680)
Can robots unravel the mystery of why bacteria are beating our best antibiotics?
Oxford University researchers are building AI-powered machines to monitor the superbug infections behind the ‘silent pandemic’
![Jessica James, second year PhD student, analyses the behaviour of bacteria using photomicrography and AI](/https/www.telegraph.co.uk/content/dam/global-health/2024/07/12/TELEMMGLPICT000385337986_17207962320720_trans_NvBQzQNjv4BqQTAoG9Ka6ECxCVPT5zp0r7ETk8P9H3T_ZmHI76ctXA8.jpeg?imwidth=680)
Walk down a few flights of stairs in a grey building in the centre of Oxford and you’ll find it – an underground lab where a group of researchers work away amidst a myriad of wires, test tubes, and computer screens.
To the untrained eye, the room, littered with beakers and pipettes, doesn’t appear that different from a GCSE chemistry classroom.
But the team here – consisting of about 20 researchers, engineers, software developers, and PhD students – are actually part of a US government-funded research project, named ‘Defeating Antibiotic Resistance through Transformative Solutions’ (DARTS).
Their mission is to develop new technologies to diagnose and treat antimicrobial resistance (AMR) – a phenomenon where pathogens evolve to evade the drugs designed to kill them.
Termed the ‘silent pandemic’, AMR currently claims 1.2 million lives annually, with projections suggesting this figure could rise to 10 million by 2050, equalling the death toll from cancer and surpassing those from malaria, HIV, and tuberculosis combined.
![Dr Harrison Steel, Associate Professor of Engineering Science at the University of Oxford](/https/www.telegraph.co.uk/content/dam/global-health/2024/07/12/TELEMMGLPICT000385338190_17207968952930_trans_NvBQzQNjv4BqqVzuuqpFlyLIwiB6NTmJwfSVWeZ_vEN7c6bHu2jJnT8.jpeg?imwidth=350)
Led by engineer Dr Harrison Steel, the University of Oxford research group has been tasked with building AI-powered machines to monitor drug-resistant infections in critically ill patients and determine effective treatments.
They have developed two systems – one named the Evolution Machine and another yet to be christened – to combat AMR.
Currently, these machines resemble a tangle of wires and tubes – a Heath Robinson aesthetic – but the goal is to refine them into affordable, accessible devices that can be widely used in hospitals and clinics worldwide to help tackle one of the most pressing health issues of our age. Here’s how they work.
The first step is diagnosis. A blood sample is taken from an unwell patient, and inserted into the diagnostic machine to quickly figure out what bacterial infection the person has contracted.
In normal clinical settings, samples are cultivated for days to allow enough bacteria to grow for human analysis, confirming the diagnosis. But by that time, it is often too late – conditions like sepsis require early intervention for successful treatment.
In the Oxford lab, there is no real wait. Within minutes, the software can identify the bacteria infecting the patient from a blood sample.
![Dr Idris Kempf, control engineer and postdoctoral researcher, who wrote the software for the AI-linked photomicrography](/https/www.telegraph.co.uk/content/dam/global-health/2024/07/12/TELEMMGLPICT000385338056_17207963584580_trans_NvBQzQNjv4BqySkuuxUK4LTxT1WX70dVCTscIxxVPcyo6nQo8OGWfGo.jpeg?imwidth=350)
Currently, the samples aren’t from actual patients but are purchased from labs, with E. coli being used as the test bacteria.
The theory, though, is that the method could work for all sorts of bacteria, including MRSA and C. diff – common hospital ‘superbugs’.
But the machines don’t stop there. Once they have identified the bug, technicians move the samples to the second system – the Evolution Machine – and feed in different antibiotics to see which will effectively kill it off and curb the infection.
An AI then projects the early results forward, meaning a technician knows exactly how to treat the bacteria in question – completing the entire process in less than half an hour.
“In really serious cases of sepsis, there’s a five to 10 per cent increase in mortality every hour – so you need something really quick. If someone has a drug-resistant infection, there’s no time to try different options.
“With our robotics, you can test different drugs very quickly and see how they will affect the infection,” explained Dr Steel.
That’s the theory, at least – but it’s not a silver bullet.
![Lab equipment for maintaining bacterial cultures](/https/www.telegraph.co.uk/content/dam/global-health/2024/07/12/TELEMMGLPICT000385338062_17207963264990_trans_NvBQzQNjv4BqqVzuuqpFlyLIwiB6NTmJwfSVWeZ_vEN7c6bHu2jJnT8.jpeg?imwidth=350)
Experts emphasise that the real issue is the dwindling effectiveness of current antibiotics, making the development of new drugs to combat bacterial infections the most urgent priority in addressing AMR.
Currently, just 27 new antibiotics for the most threatening infections are in the clinical trial stage of drug development – to compare, there were more than 1,300 cancer drugs in clinical trials in 2020.
Out of those, the WHO considers only six of them innovative enough to overcome antibiotic resistance, and just two capable of targeting the most resistant bacteria.
Without a robust pipeline of new antibiotics, even the best diagnostic tools cannot overcome the fundamental issue of dwindling treatment options in the face of rising resistance.
“New antibiotics are not the only answer to solve AMR, however” argued Prof Christopher Schofield, Director of Chemistry, Ineos Oxford Institute for Antimicrobial Research.
“Whilst we need new drugs, without mechanisms to preserve them, importantly diagnostics tests that will help doctors prescribe the right antibiotics at the right time, bacteria will inevitably become resistant to them,” said Prof Schofield, adding that innovations like this have the “potential to revolutionise the treatment of bacterial infections.”
![Colonies of MRSA bacteria on blood agar plate](/https/www.telegraph.co.uk/content/dam/global-health/2024/07/12/TELEMMGLPICT000361508220_17207969519630_trans_NvBQzQNjv4BqqVzuuqpFlyLIwiB6NTmJwfSVWeZ_vEN7c6bHu2jJnT8.jpeg?imwidth=350)
The machines are still in early development, but plans are underway to ship three of them to hospitals in the US for clinical trials in partnership with Harvard Medical School, starting early next year.
There, they will be tested using real patient samples, with the eventual goal of developing the tech into something a medical professional could run themselves, with limited expertise, in any hospital or clinic.
Typically, it takes about ten years for a new treatment to progress from initial discovery to launch, but Dr Steel believes the DARTS inventions will be ready for roll-out within the next three years.
The next issue is cost. Dr Steel doesn’t have exact numbers, but estimates the devices will cost tens of thousands of pounds to procure.
But for much of the developing world – where the burden of AMR is particularly high – the high price tag means procuring technology like this simply won’t be feasible for years to come.
“We need to ensure healthcare equity across the world so that innovative technology such as this is accessible in every country, not just high-income countries like the UK,” said Prof Schofield.
“It’s vital not only to have robust diagnostic methods as lives are on the line, but also cost-effective solutions for cash-starved healthcare systems, including in low-and-middle-income countries that currently have the highest burden of antibiotic-resistant infections,” he added.
Protect yourself and your family by learning more about Global Health Security