原文：Alterations in the gut microbiome and metabolism with coronary artery disease severity
background
Coronary artery disease (CAD) and is closely related to changes in the intestinal flora. The intestinal bacteria dominant marker metabolites are many cardiac events. However, the relationship between intestinal bacteria and CAD staging has been a mystery.
result
1. Sample Characteristics
of the 201 participants were divided into four groups, control, The SCAD group, the UA group, MI group. Statistics and very detailed clinical information. Such as diet, lifestyle and so on. And compared to the normal group, the experimental group disorders, lipid metabolism increased inflammatory reaction.
The authors used detailed statistical analysis propped up this section, experimental results much repeat. Statistical methods are as follows:
1) if four groups are normally distributed, the difference analysis to compare groups with one-way ANOVA.
2) If the four sets of data do not meet the normal distribution, the use of test KW H
3) If the two groups are normally distributed, with T-test
4) If the two are not normally distributed, the U-test with KW
. 5) for categorical variables chi-square test
6) such expression more intuitive:
... the aP <0.05 for equality BETWEEN the SCAD FC vs. Control bP <0.05 for the UA FC vs. Control equality BETWEEN cP <0.05 Control FC vs. MI for equality BETWEEN dP <<0.05 for equality between UA vs. MI. 0.05 for equality between SCAD vs. MI. eP
change between the subgroup 2. coronary artery disease
OF metabolites with serum, we were made polar (P group) and a nonpolar (L group) groups. After played metabolite, the correlation analysis of the relative expression levels do. (A) and FIG phenotype is, (b) and FIG. Is a risk factor (risk factor indicators). (C) the inter-group variation FIG greater metabolites (KW test) Videos pick out a boxplot.
3. The plurality of sets of graphics network reveal the relationship between intestinal bacteria and metabolites Serum CAD
Authors explore the relationship between intestinal bacteria and metabolites, 9 and 14 intestinal bacteria metabolic-related modules.
and FIG intestinal bacteria showed a greatest correlation host factors change. bar colors represent metadata categories. FIG Z b represents the value of the co-abundance group. C is a correlation (WGCNA network) between groups CGA
More specifically, will metabolomics by this picture of intestinal microbes, and phenotypic information linked together. (In fact, correlation analysis, red for positive correlation, blue for negative correlation ......)
4. Machine Learning add new points
OF random forest, between the groups (Con VS CAD, Con VS SCAD, SCAD VS UA, SCAD VS ACS, UA VS MI) are classified to identify the specific parameters obtained see Figure B
+ AUC I have to say machine learning curve is to add an article points ah ~
Reproduced in: https: //www.jianshu.com/p/e291250595b9