Explore the changes of intestinal flora in patients with coronavirus disease 2019 based on bioinformatics

BACKGROUND: Studies have revealed that there were significant changes in intestinal flora composition in patients with coronavirus disease 2019 (COVID-19) compared to non-COVID-19 patients, regardless of whether they were treated with medication. Therefore, a comprehensive study of the intestinal flora of COVID-19 patients is needed to further understand the mechanisms of COVID-19 development. METHODS: In total, 20 healthy samples and 20 COVID-19 samples were collected in this study. Firstly, alpha diversity and beta diversity were analyzed to assess whether there were difference in species richness and diversity as well as species composition between COVID-19 and control groups. The observed features index, Evenness index, PD index, and Shannon index were utilized to measure alpha diversity. The principal coordinates analysis (PCoA) and non-metric multidimensional scaling (NMDS) were performed to analyzed beta diversity. Linear discriminant analysis Effect Size (LEfSe) was utilized to analyze the variability in the abundance of bacterial taxa from different classification levels. The random forest (RF), Least absolute shrinkage and selection operator (LASSO), and univariate logistic regression were utilized to identify key Amplicon Sequence Variant (ASVs). Finally, the relevant networks of bacterial taxa were created in COVID-19 and control groups, separately. RESULTS: There were more species in the control group than in COVID-19 group. The observed features index, Shannon index, and Evenness index in the control groups were markedly higher than in the COVID-19 group. Therefore, there were marked variations in bacterial taxa composition between the COVID-19 and control groups. The nine bacterial taxa were significantly more abundant in the COVID-19 group, such as g-Streptococcus, f-Streptococcaceae, o-Lactobacillales, c-Bacilli and so on. In the control group, 26 bacterial taxa were significantly more abundant, such as c-Clostrjdia, o-Oscillospirales, f-Ruminococcaceae, etc. The 5 key ASVs were obtained through taking the intersection of the characteristic ASVs obtained by the three algorithms, namely ASV6, ASV53, ASV92, ASV96, and ASV105, which had diagnostic value for COVID-19. The relevance network in the control group was more complex compared to the COVID-19 group. CONCLUSION: Our findings provide five key ASVs for diagnosis of COVID-19, providing a scientific reference for further studies of COVID-19.