Applying machine learning to analyse gut bacteria could screen people for cardiovascular disease, scientists say. 

By Sofia Popov

Increasingly, research has pointed towards a strong link between our cardiovascular health and gut microbiota. Even if individual variability is high, we’ve been successfully able to differentiate between health and disease in a variety of illnesses by delving into the composition of our gut bacteria. But, can our gut microbiome be used to diagnose cardiovascular disease (CVD)? Researchers decided to find out. 

A Unique CVD Microbial Signature

CVD is a broad term that includes a range of morbid conditions: from hypertension and heart failure to atherosclerosis. As a result, the underlying mechanisms involved vary widely. Nevertheless, researchers wanted to know whether there are early warning signs that can be tracked across all these clinical conditions.

Using data from the American Gut Project, the team analysed stool samples using machine learning. Notably, they used a relatively large sample size (478 CVD and 473 non-CVD human subjects) and included a diverse pool of people.

Importantly, the model was able to identify different bacterial clusters that could then potentially help differentiate between people with and without CVD.  

Among the bacteria identified, certain bacteria were more common for each group:

  • Bacteroides, Subdoligranulum, Clostridium, Megasphaera, Eubacterium, Veillonella, Acidaminococcus and Listeria were more abundant in those with CVD.

  • Faecalibacterium, Ruminococcus, Proteus, Lachnospira, Brevundimonas, Alistipes and Neisseria were more abundant in those without CVD. 

Ultimately, the researchers were able to not only detect distinct microbial signatures, but also use this gut microbiome data as training modules for supervised ML modelling. In this way, they differentiated between these 2 groups with a promising predictive diagnostics potential. 

Overall, this study was the first to identify dysbiosis of the gut microbiota in CVD patients as a group, and apply this knowledge to develop a gut microbiome–based ML approach for diagnostic screening of CVD. This has powerful therapeutic potential moving forward. 


Reference

  1. Sachin Aryal et al. Machine Learning Strategy for Gut Microbiome-Based Diagnostic Screening of Cardiovascular Disease, Hypertension (2020). DOI: 10.1161/HYPERTENSIONAHA.120.15885