Software Can Identify Drug-Resistant Genes in Bacteria That Cause Infectious Diseases

The machine learning algorithm uses patterns in AMR proteins to detect genes in bacteria.

A medical illustration of Clostridium difficile bacteria. (Photo courtesy of the Centers for Disease Control and Prevention.)

A medical illustration of Clostridium difficile bacteria. (Photo courtesy of the Centers for Disease Control and Prevention.)

A new software program is now capable of easily identifying the presence of deadly antimicrobial-resistant genes of bacteria in the environment. These microbes are responsible for causing severe cases of pneumonia and bloodstream infections in the U.S., with over 2.8 million cases and 35,000 deaths already to date. The program’s machine learning algorithm was patterned according to the features of antimicrobial resistance (AMR), which occurs when bacteria or other microorganisms evolve or acquire genes that encode drug-resistance mechanisms. It’s this strain that can make it increasingly challenging to treat certain diseases and infections.

Researchers are particularly interested in how microbes that live in soil and water can spread and affect human health. With large-scale genetic sequencing becoming easier to utilize, AMR genes can now be easily examined as well.

The team from Washington State University (WSU) opted not to focus on sequence similarity. Instead, they observed the interactions of several features of genetic material such as structure, physiochemical and composition properties of protein sequences. This included species of Clostridium, Enterococcus, Staphylococcus, Streptococcus, and Listeria—prime causes of numerous major infections and infectious diseases such as staph infections, food poisoning, pneumonia, and life-threatening colitis due to C. difficile. The program was able to successfully classify resistant genes with an accuracy of 90 percent.

“Our software can be employed to analyze metagenomic data in greater depth than would be achieved by simple sequence matching algorithms,”said Abu Sayed Chowdhury, a PhD computer science graduate of WSU.

Shira Broschat, a member of WSU’s School of Electrical Engineering and Computer Science, also added: “The virtue of this program is that we can actually detect AMR in newly sequenced genomes. It’s a way of identifying AMR genes and their prevalence that might not otherwise have been found. That’s really important.”

According to the researchers, game theory also played a significant role in forming their strategic framework.

The software package is available for download and for use by other researchers who are interested in locating AMR in large pools of genetic material. The team expressed that researchers can also retrain the algorithm to improve the software as more data and sequences become available.

The complete study can be found in Scientific Reports.

For more news and stories, check out how this plastic wrap repels bacteria here.