Artificial Intelligence on its way to help to predict drug resistant bacteria

Artificial Intelligence on its way to help to predict drug resistant bacteria

Researchers of Washington state university have developed a feasible software in their lab which helps to identify drug-resistant contributing genes in bacteria.

The software developed makes it feasible to use and easily identify the fatal anti-microbial resistant bacteria existing with us in nature. Figures show that anti-microbial resistance bacteria alone cause more than 2.8 million cases of deadly pneumonia, blood infection, and 35000 deaths annually. Researcher Abu Sayed- Doctoral graduate in computer science, Shira Broschat from School of Electrical Engineering and computer science, and Douglas Call from Paul G. Allen School for Global Animal Health have shared the insight of their work with the Scientific Reports journal.

Antimicrobial resistance is a process that is observed when a microbe acquires or evolves with a drug resistance gene thus further causing AMR. Bacteria responsible for staphylococcus or streptococcus infection or in that manner tuberculosis and pneumonia, all these infectious diseases have developed due to the occurrence of drug resistance strains thus causing hindrance in the treatment regime. The AMR issue is estimated to worsen more in the coming decades in terms of AMR-related infection, deaths, and an increase in health costs as the bacteria is evolving in a way where there is a limited option of antibiotic treatment against it.

“We need to develop tools to easily and efficiently predict antimicrobial resistance that increasingly threatens health and livelihoods around the world,” said Chowdhury, lead author of the paper.

With the ease of large-scale genetic sequencing, scientists are looking in the environment for the presence of AMR genes. They are interested to know the habitat of such genes and how potentially such AMR microbes can spread and affect human health. While they can identify genes that are similar to known AMR-resistant genes, they are probably missing genes for resistance that look very unique from a protein sequence perspective.

The team at WSU developed a machine learning algorithm that has a unique feature of having data of AMR proteins instead of using gene sequences to find a similarity and identify AMR genes. The above algorithm was developed with the help of a theory known as Game theory. Game theory is a tool used in various fields more precisely in economics, to model strategic interactions between game players, here, in this case, helps to identity AMR genes. Using algorithm and theory together, the researcher looked at the interaction of various aspects such as genetic material, structure and physicochemical properties, and properties of protein sequences rather than simply looking at sequence similarity.

“Our software can be employed to analyze metagenomic data in greater depth than would be achieved by simple sequence matching algorithms,” Chowdhury said. He further added, “This can be an important tool to identify novel antimicrobial resistance genes that eventually could become clinically important.”

“The virtue of this program is that we can actually detect AMR in newly sequenced genomes,” Broschat said. “It’s a way of identifying AMR genes and their prevalence that might not otherwise have been found. That’s really important.”

The pioneer research team at WSU took Clostridium, Enterococcus, Staphylococcus, Streptococcus, and Listeria species where resistance genes are found. The above-mentioned bacteria are one of the major causes of infections such as staph infection, food poising, life-threatening pneumonia, and colitis. The algorithm developed by the team was able to categories the resistance genes accurately up to 90 percent.

The algorithm is embedded as a software package that is downloadable easily and can also be used by the scientific fraternity to look into AMR in a large pool of genetic material. The software can be updated timely. As more sequences and data become available, researchers will be able to continue with the algorithm without any hassle. “You can bootstrap and improve the software as more positive data becomes available,” Broschat said.