Skip to content Skip to footer

Using artificial intelligence in the form of deep learning, researchers from MIT have discovered compounds capable of killing methicillin-resistant Staphylococcus aureus (MRSA), a drug-resistant bacteria that reportedly causes over 10,000 deaths in the U.S. each year. This breakthrough was achieved by training a deep learning model using predictive information based on antibiotic potency of a large selection of compounds. This information was then used to identify which compounds may contain this potency.

MRSA, which infects more than 80,000 people in the U.S. every year, often causes skin infections and pneumonia and in severe cases, can lead to sepsis, a fatal bloodstream infection. With the help of this advanced technology, the scientists were successful in identifying potential antibiotic drugs that can be used against various drug-resistant bacteria.

Typically, the models used for this type of prediction are considered to be black boxes as there is it’s unclear what basis the model uses to make its predictions. In an attempt to make these predictions more explainable, the researchers adapted an algorithm known as the Monte Carlo tree search which had been used in other deep learning models, such as AlphaGo. This resulted in the model being able to predict antimicrobial activity as well as what contributes to its activity within the molecule.

To focus on potential drugs, the team worked with three additional deep learning models to forecast the toxicity of the compounds on different types of human cells. Using these models, the team was able to screen about 12 million commercially available compounds to identify effective antibiotics that also had minimal adverse effects on the human body.

In mouse models, the selected compounds were able to reduce the MRSA population by 10 times. The compounds appear to kill bacteria by inhibiting their ability to maintain an electrochemical gradient across cell membranes, necessary to produce ATP, molecules that store energy in cells. The team’s findings have been given to Phare Bio, a non-profit launched by Collins, for further analysis of the compound’s potential clinical usage. The MIT researchers are also striving to design more potent drug candidates based on their findings.

Leave a comment

0.0/5