JAKARTA - Artificial intelligence or AI helps scientists find a promising new class of antibiotics to fight MRSA, a drug-resistant bacteria that can cause serious infections.
Euronews, quoted on Monday, July 27, reported that the findings came from research published in the journal Nature. The study was co-authored by 21 researchers.
This research focuses on methicillin-resistant Staphylococcus aureus or MRSA. This bacterium can cause mild skin infections to severe diseases, such as pneumonia and bloodstream infections.
According to the European Centre for Disease Prevention and Control or ECDC, almost 150,000 MRSA infections occur each year in the European Union. In the same region, almost 35,000 people die each year from antimicrobial resistant infections.
Antimicrobial resistance occurs when germs, including bacteria, are no longer susceptible to previously effective drugs. This condition makes infections more difficult to treat.
A team from the Massachusetts Institute of Technology or MIT used a deep learning model to search for compounds that could potentially be antibiotics. Deep learning is a part of AI that mimics the way neural networks work to read patterns in data.
James Collins, professor of Medical Engineering and Science at MIT and one of the study's authors, said the advantage of this research is in the AI model that is easier to read how it works.
"The important takeaway is that we can look at what the model learned to make predictions that certain molecules will be good antibiotics," Collins said.
"Our work provides a time-efficient, resource-efficient framework and gives an understanding of the mechanism from a chemical structure point of view, in a way that we have not had until now," he said.
The researchers first evaluated about 39,000 compounds to see their antibiotic activity against MRSA. The data from the test results, along with information on the chemical structure of the compounds, was then entered into an AI model.
Felix Wong, a postdoctoral researcher at MIT and Harvard and one of the lead authors of the study, said his team wanted to open up what has been called a "black box" in AI models.
"What we want to do in this study is open the black box. These models consist of a very large number of calculations that mimic neural connections, and no one really knows what's going on behind them," said Wong.
To tighten the search, researchers used three additional deep learning models. These models are trained to assess the toxicity of compounds in three types of human cells. Toxicity is the level of a substance's likelihood of having a harmful effect on the body.
After looking at the ability to kill bacteria and the level of danger for human cells, researchers chose the most promising compounds.
Around 12 million commercially available compounds were then screened using the AI model series.
The model found compounds from five different classes that were predicted to be active against MRSA. The classes were distinguished based on specific parts of the chemical structure within the molecule.
The researchers then obtained about 280 compounds and tested them against MRSA in the laboratory. From there, they found two promising antibiotic candidates from the same class.
Antibiotic candidates are compounds that still need to be tested further before they can become drugs.
Euronews said this finding is still at the research stage. In tests on two mouse models, one for MRSA skin infection and another for systemic MRSA infection, each compound reduced the MRSA population by up to 10 times.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
