The observed species (Sobs) index can be used to judge the richness of microbial colonies, while the Shannon index can be used to judge the diversity of such colonies. The number of microorganisms in the large intestines of mice that were gagged with normal saline is what "Group C" means. After the gavage of antibiotics, the diversity of organisms was greatest in Group M. As can be seen, compared with Group C, the species diversity and richness of the intestinal microbial community in Group M were lower. The T group reveals that specific bacteria exhibited a pattern of recovery and expansion (Figs. 2 and 3). After administration of A. lancea extract, the diversity and abundance of microorganisms in the large intestines of mice tended to recover and grow.

Diversity index. The abscissa is the sample, and the ordinate is the number of species observed at the genus level

Student’s t test histogram of intergroup differences. The abscissa is the group name, and the ordinate is the mean value of Shannon index at the OUT classification level. (*P < 0.05, **P < 0.01, ***P < 0.001)

The bar diagram in Fig. 4 indicates the compositions of numerous phyla at various taxonomic levels, as well as the corresponding dominant phyla and their proportions. At the phylum level, the biggest proportion of bacteria among the three groups was Firmicutes. The proportion of Bacteroides reduced considerably after antibiotic gavage, and Proteobacteria replaced it as the dominating phylum. Following treatment with A. lancea extract, Bacteroides and Actinomycetes proportions increased. When antibiotics were administered to mice, the phyla of bacterial in their intestines changed. The percentage of probiotics decreased. The equilibrium of flora was upset, and the usual growth of helpful beneficial bacteria was impeded. However, after treatment with A. lancea extract, the number of beneficial bacteria began to recover, and the proportion of harmful bacteria tended to return to its level prior to antibiotic administration, indicating that A. lancea extract had a significant effect on restoring the balance of intestinal flora in mice.

Histogram of sample community composition. The abscissa is the proportion of species, and the ordinate is the sample number

For heatmap analysis, the 50 genera with the highest abundance were screened. The genera Polyformis, Prevosiella, and Muribaculaceae were more abundant in the normal group, as depicted by the community heatmap at the genus level (Fig. 5). Antibiotics decreased the number of bacteria in the normal group, but Pseudomonas aeruginosa, Enterococcus, and Clostridium increased in number. The leading colonies were Escherichia Shigella and Otherbacter, and several of the more abundant communities in the normal group were recovering. However, the abundance of some medications was lower than that of antibiotics (Group A), which may have been due to the brief time of oral A. lancea extract administration and the ease with which other microorganisms were recovered from the environment formed by A. lancea. In contrast, several bacteria were more sensitive to the environment generated by antibiotics or A. lancea..

Heatmap of sample community. The abscissa is the sample number, the ordinate is the species name, and the legend is the species abundance value

We utilized a Circos diagram (Fig. 6) to examine the composite proportions of dominant phyla in various samples. Under normal conditions, Firmicutes and Bacteroidetes are the dominant groups. In mice treated with antibiotics, Proteobacteria grew from 3.8% to 32%, whereas Bacteroidetes declined from 43% to 5.5%, demonstrating that antibiotics can alter the composition of the intestinal flora and hence produce an intestinal-flora problem. After i.g. treatment with A. lancea extract, Mycorrhizae and Bacteroidetes were once again in the dominating floras. The drop in Proteobacteria and significant rise in Bacteroidetes (from 5.5% to 13%) demonstrated that A. lancea extract might enhance the number of beneficial bacteria in the digestive tract and restore the balance of intestinal flora.

Circos diagram. In the Circos sample and species diagram, the large semicircle (right half circle) represents the distribution proportion of a species in different samples under a certain classification level (outer ribbon: species, inner ribbon color: different groups), and the length represents the specific distribution proportion; The small semicircle (left half circle) represents the composition of different species in a sample (outer band: grouping, inner band: species), and the relative abundance specific to length

We used the Kruskal–Wallis H test (Fig. 7) to assess statistically differences between groups and to determine differences in species richness within each group. As shown in Fig. 7, the three groupings included Muribaculaceae, Clostridium, Lactobacillus, Prevotella (P ≤ 0.001), Eschia, P. aeruginosa, multiform rod-shaped bacteria, and Romboutsia (0.001 < P ≤ 0.01). There were significant differences between the eight different genera. In addition, the color columns of the three groups indicated that the species richness of Group C, which was initially low, increased after the administration of antibiotics. Examples include P. aeruginosa and Clostridium. These findings demonstrated that antibiotics can alter the equilibrium of intestinal flora. Following treatment with A. lancea extract, the relative abundance of these two species fell on average. These results demonstrated that A. lancea extract may restore the average relative abundance of intestinal microflora and treat intestinal dysbiosis in mice.

Multi-group comparison of significant differences between groups. The left column shows the species name under the genus classification, and the corresponding column shows the average relative abundance of the species in each sample group. Different groups have different colors. The right-hand side is the p-value. (*P < 0.05, **P < 0.01, ***P < 0.001)

In addition, PCA based on OTUs indicated that these populations had distinct microbiome characteristics. As represented in Fig. 8, the results from the three groups showed a distinct difference phenomenon, confirming the accuracy of our experimental methodology. At the same time, mouse A6 in Group A showed obvious separation from other members of the same group, which have been attributable to unique causes.

Principal component analysis. The closer the two sample points are, the more similar the species composition of the two samples is

Significant separation was also observed in the hierarchical-clustering results (Fig. 9), which was consistent with PCA results. Meanwhile, the three classifications shown in Fig. 9 are unique, indicating that the test technique and grouping strategy we picked were appropriate. The performance of test objects in subgroups can be seen from the data set of Group C and Group A_A (Group T), but mouse A6 in Group A is clustered distant from the other members of the same group in the PCA graph due to its individual differences.

Significance of differences between groups. Clustering tree graph for multi-group comparisons between samples

The abovementioned analytical results showed that the methodology utilized in our investigation was precise and efficient. In addition, they also proved that antibiotics threw mice's richness and diversity of intestinal flora out of balance, resulting in intestinal-flora instability. The intestinal-flora structure of mice treated with A. lancea extract was considerably different from that of animals treated with antibiotics alone. Diversity and abundance of mouse intestinal flora tended to recover, as did the intestinal flora of mice not administered antibiotic injections. Using metabolomics, the impact and implications of these changes will continue to be examined and explored.

Correlation heatmaps were obtained using Pearson’s correlation coefficient (PCC) and Euclidean distance algorithms in both anionic and cationic modes (Fig. 10). As seen in Fig. 10 and Table 2, there were differences between Groups C and M, particularly in anionic mode. There were also changes between Groups C and T, but these were very minor in cationic mode, and differences were significant in anionic mode.

Sample correlation heatmap. A Sample correlation heat map in C vs M positive ion mode; B Sample correlation heat map in C vs M negative ion mode; C Sample correlation heat map in C vs T cationic mode; D Sample correlation heat map in C vs T anionic mode

As shown in the PCA score chart (Fig. 11), we noticed a considerable separation between Group M and other groups in both cationic and anionic modes. In Fig. 11D, the PCA score chart of the T versus (vs.) M comparison group in cationic mode revealed a partial overlap between Groups T and M, probably because of two-dimensional nature of the exhibited images. Group M was separated from remaining groups.

PCA score chart. A PCA score of comparison group M vs C in anionic mode; B PCA score of M vs C comparison group in cation mode; C PCA score of comparison group T vs M in anionic mode; D PCA score of T vs M comparison group in cation mode

In addition, additional OPLS-DA model analysis (Fig. 12) also indicated that Groups M and T were distinct, showing that A. lancea extract significantly changed the physiological-metabolic condition of mice following antibiotic therapy.

OPLS-DA score chart. A opls-da score of M vs T comparison group in anionic mode; B pls-da score of T vs M comparison group in cation mode

To verify the validity of the OPLS-DA model, we employed response permutation testing to assess its precision (Fig. 13). The main parameters were as follows: Anionic mode: R2X (cum) = 0.612, R2Y (cum) = 0.989, Q2 (cum) = 0.827; Cationic mode: R2X (cum) = 0.538, R2Y (cum) = 0.994, Q2 (cum) = 0.808. The above parameters are all more than 0.5, R2Y had a high value. This investigation produced a model with great precision, stability, and dependability.

Response permutation test chart. A Response permutation test in anion mode; B Response permutation test in cation mode

Using a volcano map (Fig. 14), we were able to quickly determine the statistical significance of variations in metabolic-expression levels between antibiotics and A. lancea extract after gavage and the statistical significance thereof. Figure 14A/14C depicts a multitude of substantially up-regulated regions. Groups T and C differ significantly in up- and down-regulated regions (red and green, respectively) as depicted in Fig. 14C. However, because there is only one value in the red area beyond abscissa 4, this result may have been caused by previously indicated. In addition, the kind of bacteria after Atractylodes lavage was superior to that after antibiotic lavage, and the number of beneficial bacteria was inconsistent with that before lavage. The majority of strongly upregulated and downregulated regions fall within the range of abscissa 1 to 2, with significantly upregulated regions comprising a higher proportion.

Difference volcano map. A M vs C differential volcanic map; B T vs M differential volcanic map; C T vs C differential volcanic map

The top 30 significantly differential metabolites in the comparison groups (C vs. M and M vs. T; Tables 3 and 4) was identified using a screening condition of P < 0.05, VIP_pred_oPLS-DA > 1, fold change [FC] > 1, or FC < 1. Compared to Group C, the contents of phospholipids such as cephradine, theophylline, lysophosphatidylcholine (LysoPC), and lysophosphatidylethanolamine in Group M were significantly decreased, while the concentrations of 2-phenylglycolic acid, cinnamyl glycine, catechol sulfate, gibanoic acid M, 3-indolepropionic acid, and 5-oxy-ferulic melanin in Group M were significantly increased. Compared to Group M, lipids such as phosphatidylcholine and terpinyl anthranilate were increased in Group T, but emodin, hydroxyl hexadecarboxylic acid, 3-carboxylic 2,3,4,9-tetrahydro1h-pyridine [3,4-B] indole-1-propionic acid, daidzein, and 2-hydroxyundecanoic acid were significantly decreased, indicating that metabolites were significantly changed after antibiotic treatment. A. lancea extract in moderation and the number of beneficial bacteria gradually recovered after i.g. administration.

The cluster analysis of the first thirty metabolites is depicted in Fig. 15. As shown, the expression levels of LysoPC, lysophosphatidylethanolamine, arachidonic acid (AA), and other substances were significantly altered after treatment. The number of metabolites annotated to the lipid metabolism pathway was the greatest, followed by those annotated to the amino acid metabolism pathway (Fig. 16). Differentially annotated lipid metabolites comprise AA, taurocholic acid, traumatic acid, palmityl l-carnitine, dodecanedioic acid, 21-deoxycortisol, 3-oxygen-sulfolactose ceramide, 17-hydroxyprogesterone, 13S-hydroxyoctadecenoic acid, phosphatidylcholine (24:1 [15Z]), phosphatidylcholine (18:3 [6Z, 9Z, 12Z]), and phosphatidylcholine (16:0). The amino acid metabolic pathway were such differential metabolites as phenylacetylglycine, L-methionine, indoleacetaldehyde, thyroxine, gin, and m-coumaric acid.

Group M vs Group C cluster analysis of the first 30 metabolites in the control group

KEGG pathway charts. A Group M vs Group C; B Group T vs Group M

Figure 17 depicts the KEGG enrichment analysis of differential metabolites as a diagram of bubbles. As shown in the enrichment analysis bubble diagram of Group M vs Group C, the bubbles in the bile secretion pathway are the greatest, as was the number of metabolites enriched to metabolic concentration (4). The second pathway was phenylalanine metabolism, which was enriched with three metabolites. Two metabolites were enriched for the linoleic acid metabolism, the α-linolenic acid metabolism, the glycerol phospholipid metabolism, the mutual conversion of pentose and glucuronic acid, the arginine metabolism, and the proline metabolism. As seen in the enrichment analysis bubble diagram for Group T vs. Group M, the bubbles in linoleic acid metabolism, phenylalanine metabolism, tryptophan metabolism, steroid biosynthesis, bile secretion, ovarian-steroid production, and other pathways are the largest and identical in size, indicating that the number of metabolites enriched to metabolic concentration was equivalent. The ovarian-steroid route has the highest degree of enrichment, followed by linoleic acid metabolism.

KEGG enrichment analysis. A Group M vs Group C; B Group T vs Group M

According to the KEGG topology analysis bubble diagram of differential metabolites (Fig. 18), the most important KEGG pathway was MAP00590, namely, the AA metabolism pathway. We used Interactive Pathways Explorer (iPath) v3.0 for visual examination of the metabolic pathways of all differential metabolites (Fig. 19). As shown in Fig. 18, lipid metabolism and amino acid metabolism accounted for the majority.

KEGG topology analysis. A KEGG Topology Analysis of M vs C; B KEGG Topology Analysis of T vs M

iPath metabolic pathway. A iPath metabolic pathway T vs M; B iPath metabolic pathway M vs C; iPath metabolic pathway of T vs C

To determine the potential relationship between changes in intestinal flora in feces and changes in metabolites in serum from mice, we compared three conditions and analyzed correlations between differential metabolites in serum and intestinal microflora using Spearman's correlation coefficient (SCC). Figure 20A shows that G_Odoribacter, G_Gordonibacter, and Helicobacter were positively correlated with phenylacetylglycine, lentialexin, and LysoPC. Staphylococcaceae and Streptococcaceae were positively correlated with phenyluronic acid, lentialexin, and LysoPC, as depicted in Fig. 20B. Helicobacteraceae showed a positive correlation with the A group. In the C group, Odoribacter was positively correlated with phenylacetylglycine, lentialexin, and LysoPC, just as it was in the A group. G_Ruminococcaceae was also positively correlated with these three metabolites, as well as with other species of G_Ruminococcaceae. As depicted in Fig. 20, the connection between bacterial species and metabolites in the three groups was broadly comparable, however species differences led to variances in the final results. As a result, we hypothesized that this link was highly dependable and that the metabolites in question could be those of these species or derivatives thereof.

Spearman’s correlation coefficient. A M is associated with the C analysis, B T is associated with the C analysis, and C T is associated with the M analysis