Potential biomarkers of acute myocardial infarction based on the composition of the blood microbiome

It is difficult to distinguish between acute myocardial infarction (AMI) and unstable angina (UA) due to their similar clinical features. In recent years, studies have shown that microbiomes have great potential in distinguishing diseases. The purpose of this study is to describe the composition of serum microbiome in the AMI and UA by 16S rDNA sequencing. Based on the high-throughput detection platform and 16S rDNA amplification sequencing technology, this study detected the blood microbial composition of 55 patients with AMI and 62 patients with UA. Alpha diversity and Beta diversity analysis were used to compare the differences in microbial composition and bacterial colony structure between AMI and UA groups. We perform PCoA (Principal Co-ordinates Analysis) based on Unweighted Unifrac distance. In addition, various statistical methods were employed to examine the significance of differences in microbial composition and genus between the two groups. PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) was employed to predict KEGG (Kyoto Encyclopedia of Genes and Genomes) function from 16S sequencing data. Random forest was applied to identify biomarkers and construct the diagnostic model. Subsequently, the stability of the model was verified by 10-fold cross and the diagnostic effectiveness was evaluated through ROC (Receiver Operating Characteristic). In this study, we found that alpha diversity index of serum microbiome in AMI group was significantly higher than in UA group. T-test analysis demonstrated that the UA group presented a higher abundance of Ralstonia, Faecalibaculum and Gammaproteobacteria, while Bacteroides, Sphingomonas, Faecalibaculum, Haemophilus, Serratia, Bifidobacterium and Chloroplast were more abundant in the AMI group. The LefSe (LDA Effect Size) analysis showed that the Gammaproteobacteria, Proteobacteria, Ralstonia pickettli, Ralstonia, Burkholderiaceae and Burkholderiales were enriched in UA group, and Bacteroidales, Bacteroidia, Bacteroidota, Clostridia and Firmicutes were more abundant in the AMI group. Ten bacterial diagnostic models were constructed in the random forest. The area under the curve (AUC) in the training set was 88.01%, and the AUC value in the test set was 95.04%. In this study, the composition of blood microorganisms in the groups of patients with AMI and UA has been analyzed, providing novel insights for understanding the pathogenesis of AMI; Blood microbiome may serve as novel diagnostic biomarkers of AMI.