Skip to main content
. 2023 Nov 9;14:7249. doi: 10.1038/s41467-023-43167-5

Fig. 2. Microbiome-driven precision diagnostics for STEMI using machine learning.

Fig. 2

a Workflow for supervised machine learning. b The performance (recorded as area under curve, AUC) of microbiota-based model with different algorithms for the diagnosis of myocardial infarction. n = 10-fold cross-validations. c Receiver operating characteristic (ROC) curve performance using Ctrl-predominant bacteria Bifidobacterium adolescentis (Ba.) and Bifidobacterium ruminantium (Br.). d ROC curve performance using STEMI-predominant bacteria include Streptococcus parasanguinis (Sp.), Streptooccus salivarius (Ss.) and Butyricimonas virosa (Bv.). ROC curves were analyzed with Wilson/Brown test with 95% confidence interval.