AI to find an antibiotic against a superbug

Source: By The Indian Express

In a major breakthrough for the use of Artificial Intelligence (AI) in the field of medicine, scientists from the United States and Canada have found a new antibiotic – powerful enough to kill a superbug – using AI.

Superbugs are bacteria that are resistant to several types of antibiotics. Each year these drug-resistant bacteria infect more than 2 million people in the US and kill at least 23,000, according to the US Centers for Disease Control and Prevention (CDC).

What is Acinetobacter baumannii?

The study (‘Deep learning-guided discovery of an antibiotic targeting Acinetobacter baumannii’) published in the journal Nature Chemical Biology on 25 May 2023 dealt with the bacterium Acinetobacter baumannii and saw participation from Canada’s McMaster University and Massachusetts Institute of Technology (MIT) in the US.

In 2017, the bacterium was identified by the World Health Organization (WHO) as one of the world’s most dangerous antibiotic-resistant bacteria. Notoriously difficult to eradicate, A. baumannii can cause pneumonia, meningitis and infect wounds, all of which can lead to death, according to the University of McMaster. A. baumanni is usually found in hospital settings, where it can survive on surfaces for long periods, it said.

The WHO’s list of superbugs highlighted bacteria that are having built-in abilities to find new ways to resist treatment and can pass along genetic material that allows other bacteria to become drug-resistant as well.

How do bacteria become resistant to drugs?

Antibiotics are medicines used to prevent and treat bacterial infections. Antibiotic resistance occurs when bacteria change in response to the use of these medicines, says the WHO. This ultimately threatens the ability of medicines to treat common infectious diseases.

Where antibiotics can be bought for human or animal use without a prescription, the emergence and spread of resistance is made worse, it says, cautioning against overconsumption of medicines without medical professionals’ recommendation for treating common illnesses.

The WHO lists infections such as pneumonia, tuberculosis, and foodborne diseases as becoming harder to treat with existing medication due to increasing anti-bacterial resistance.

How did researchers use AI in this case?

Narrowing down the right antibacterial chemicals against bacteria can be a long, difficult process. This is where algorithms come in because the concept of AI is based on the process of machines being given large amounts of data and training themselves on identifying patterns and solutions based on them.

According to MIT, the researchers first exposed A. baumannii grown in a lab dish to about 7,500 different chemical compounds, to see which ones could help pause the growth of the bacterium.

Then they fed the structure of each molecule into the machine-learning model. They also told the model whether each structure could prevent bacterial growth or not. This allowed the algorithm to learn chemical features associated with growth inhibition.

Once the model was trained, the researchers used it to analyse a set of 6,680 compounds. This analysis took less than two hours and yielded a few hundred results. Of these, the researchers chose 240 to test experimentally in the lab, focusing on compounds with structures that were different from those of existing antibiotics.

Those tests yielded nine antibiotics, including one that was very potent and effective at killing A. baumannii. This has been named abaucin.

Using AI, we can rapidly explore vast regions of chemical space, significantly increasing the chances of discovering fundamentally new antibacterial molecules,” said Jonathan Stokes, lead author of the paper and an assistant professor in McMaster’s Department of Biomedicine & Biochemistry.