A Helping Hand
Elinor Jones discusses the use of AI in the search for new antibiotics
Whilst COVID-19 is posing a highly dangerous and very current threat, a long-term healthcare crisis, often overlooked in mainstream media, has been developing in recent years: antibiotic resistance.
Antibiotic resistance, where the overuse of drugs such as penicillin causes deadly bacteria to grow despite toxic compounds aiming to kill them, was and still should be claimed as the greatest challenge of the twenty-first century , resulting in over sixty thousand deaths in England between 2017 and 2018. Compounds previously used to treat bacterial infections have become risky, with routine hospital admissions running risk of infection development that cannot be controlled by antibiotics currently in circulation.
The mid-twentieth century saw a honeymoon period for antibiotic development, with more and more compounds being identified and more infections stopped. This honeymoon period has long been over, but has made room for the emergence of a new era in medical research- the age of artificial intelligence. Scientists in the US have identified potentially life-saving antibiotics using an algorithm that can spot useful compounds much faster than human scientists, analysing one hundred million potential molecules in days. The algorithm, written by scientists at the Massachusetts Institute of Technology (MIT), was developed to seek out antibiotic compounds that kill bacteria using different mechanisms to those already on the market, hoping to flatten the curve on deaths caused by antibiotic resistance.
“The mid-twentieth century saw a honeymoon period for antibiotic development”
This algorithm, which helped identify a compound that could be used to kill up to thirty-five types of bacterial infection (some deadly) would not, however, replace the need for human scientists. In fact, decision-making usually made by the human brain inspired the algorithm , with the ability to be trained to spot compounds with the highest effectiveness against E. coli. Once candidates were identified, scientists at MIT performed physical testing on the compounds to test the feasibility of these drugs as antibiotics, with halicin being the most powerful from around one hundred candidates tested, although several other compounds are set to be tested for their potential antibiotic properties. If halicin is found highly effective in the lab through cell culture and rodent studies, we could see it progressing to human trials in the future.
Whilst the time taken from theorising the molecular structure of an antibiotic to performing tests in the lab has been shortened significantly, it could still be a while before we see new drugs on the market as rigorous safety and efficacy testing will be required. However, this on-going crisis is one-step closer to resolution with a little helping hand from artificial intelligence.
