Big Data, Machine Learning Come to Bacteriology in Group A Strep Study

NIAID Now | February 21, 2019

NIAID scientists and funding have helped colleagues at Houston Methodist Hospital establish a “layered” big-data approach to analyze microbial genetic changes that can make some bacteria extremely dangerous to people. They used a Streptococcus pyogenes bacterium (also known as strep, group A Streptococcus, or GAS) in the project, but believe the approach will work with many bacterial pathogens, and the data produced could yield new therapeutics and vaccines.

Think of how automakers use data to design and then assemble a vehicle—analyzing how small changes affect performance, durability, and safety. Then, despite the best duplication efforts on the factory line, no two vehicles are exactly alike. The same is true for the progeny of bacteria.

Like the automaker who uses data to find trends, aberrations and ways to make improvements—a type of machine learning—the scientists used their data to find and predict genetic variations at each step of the bacterial replication cycle. That information can help them determine whether no disease, mild disease or severe disease could result, and hopefully guide their disease prevention efforts.

The researchers intricately assessed more than 2,000 samples of strep isolated from patients from six countries. Strep can cause a range of human illnesses from strep throat to “flesh-eating” disease. There is no vaccine against strep and it can be difficult to treat with antibiotics.

The layers of data the scientists used included analyzing the DNA (genomics) of 2,101 samples; screening and selecting 492 of those samples for RNA analysis (transcriptomics); then screening again and choosing 50 of those samples to assess disease severity (virulence).

Assessing the relationships between multiple types of genetic analysis is new to bacteriology, though the Methodist group says the approach has been successful in studying cancer cells.

Next up, the group hopes to use its discoveries to develop new therapeutics against strep, especially strains that cause puerperal sepsis, an infectious disease of women who have given birth, also known as childbed fever.

Several research groups in Denmark, Finland, Iceland and Norway collaborated with the Methodist Hospital group and provided patient samples; Canada and the United States also provided patient samples. NIAID’s Laboratory of Bacteriology at Rocky Mountain Laboratories and Weill Cornell Medical College and Cornell University in New York also collaborated on the project.

To learn more about NIAID-supported vaccine development efforts, see Group A Streptococcus Vaccine Research. For information about diagnosis and treatment of strep throat, see CDC’s Strep Throat: All You Need to Know.

Reference

P Kachroo et al. Integrated analysis of population genomics, transcriptomics and virulence provides novel insights into Streptococcus pyogenes pathogenesis. Nature Genetics DOI: 10.1038/s41588-018-0343-1 (2019).

Content last reviewed on February 21, 2019