Clinical characteristics of the study population

Table 1 displays the demographic and clinical characteristics of eleven AP participants and fifteen control subjects. In terms of age, gender, hypertension percentage, and body mass index (BMI), there was not a statistically significant difference found between the two groups (P > 0.05). Table 1 displayed clinical characteristics such as CTSI, CRP, IL-6, WBC, PCT, D-dimer, and PLT.

Table 1 Clinical data of the patients with AP and control subjectsOTU clustering and alpha diversity analysis of gut mycobiota

Using Illumina high-throughput sequencing technology, a total of 1,916,597 raw reads were collected from 26 samples. From these, 1,486,419 high-quality sequences were selected, with an average of 57,170 sequences per sample. We produced 678 OTUs using 97% as the similarity cutoff. The Venn diagram showed that 173 of the 678 OTUs were shared by both groups, whereas 140 OTUs were unique to the test group, and 365 were specific to the control group (Fig. 1a). The microbial richness of the fecal sample reached near saturation at the sequencing depth used, as seen by the subsequent rarefaction curves. Given that the OTU-based Shannon-Wiener curve had already reached a plateau, it may be concluded that our sequencing depth was satisfactory (Fig. 1b).

Fig. 1

Alpha diversity analysis of fecal mycobiota of healthy volunteers and AP patients. (a) Venn diagram demonstrates the shared and unique OTUs in both groups. (b) Shannon diversity curve of a single sample in each group. (c-h) The index values of alpha diversity (c: Shannon index; d: Simpson index; e: Coverage index; f: Ace index; g: Sobs index; h: Chao index)

The evaluation of alpha diversity includes the following: Chao, Shannon, Ace, Simpson, Coverage, and Sobs (also known as observed species). Indicators of community variety, the Shannon and Simpson indices, showed statistically significant changes (P < 0.05). Further analysis revealed no statistically significant variation in the coverage index, a measure of community coverage. These findings suggest that the diversity of intestinal fungi in patients with AP was significantly lower compared to the control group (Fig. 1c-e). There were no significant differences in Ace, Sobs, and Chao in response to community richness (P > 0.05) (Fig. 1f-h).

Analysis of the beta diversity based on OTU levels

To examine the similarity or dissimilarity in the general community structure between the afflicted and healthy groups, the species diversity was analyzed by intergroup comparison. The principal co-ordinates analysis (PCoA) was employed to measure beta diversity, utilizing the bray_curtis distance metric. The Adonis analysis revealed a substantial distinction in the gut fungal community between the two groups (R2 = 0.1361; P = 0.0160) (Fig. 2a). The ANOSIM analysis also yielded consistent outcomes (R = 0.1664; P = 0.017) (Fig. 2b). We confirm this outcome using PCoA with both weighted and unweighted UniFrac distances (Figure S1c and d).

Fig. 2

Analysis of beta diversity based on OTU and genus levels. (a) PCoA plots of beta diversity based on Bray-Curtis distances Adonis analysis in different groups. (b) PCoA plots of beta diversity based on Bray-Curtis distances ANOSIM analysis in different groups. (c) A hierarchical clustering tree was created to show the relationships among all the samples

Moreover, by employing non-metric multidimensional scaling (NMDS) in conjunction with the bray_curtis distance to assess beta diversity, Adonis and ANOSIM analyses (R2 = 0.0909; P = 0.0250; R = 0.1664; P = 0.0170, respectively) revealed a distinction in the intestinal fungal communities of the two groups (Figure S1a and b). This finding suggests that the disparities between the groups were more pronounced than those within the groups, suggesting that the classification of the control (healthy) and test (AP) groups was significant for the purposes of this research. At the OTU level, hierarchical cluster analysis partitioned the fungi into discrete branches (Fig. 2c). The examination of colony typing in relation to the dominant colony structure of the two groups indicated a statistically significant distinction in the typing composition (Figure S1e and f).

Acute pancreatitis altered gut mycobiota taxonomic composition

When accounting for the species of all OTUs, a total of 8 phyla were found in the test group, while only 6 phyla were detected in the control group. Ascomycota and Basidiomycota dominated the fungal microbiome of both populations. The percentage of dominating flora in the test group was 92.3% and 7.201%, respectively, whereas they were 77.78% and 21.05% in the control group. The relatively average abundance of the fungal microflora at the phylum level is shown in Fig. 3a. At the class level, the numbers of fungi in Saccharomycetes (67.22%) were the largest in the test group and the number of other classes like Eurotiomycetes (21.83%), Tremellomycetes (4.00%) and Microbotryomycetes (2.53%) were less than 50%. Additionally, the dominating classes detected in the control group were Eurotiomycetes (45.30%), Saccharomycetes (23.97%), Agaricomycetes (14.54%) and Sordariomycetes (4.76%) in descending order (Fig. 3b). The composition of gut flora in both groups had a similar organization at the order and family levels, as observed at the class level. The bar charts that are specific to the community are displayed in Fig. 3c and d. At the genus level, the fungal microbiota was dominated by Aspergillus in the control group, followed by Candida, Ganoderma, Penicillium, Auricularia, and Talaromyces with proportions of 28.65%, 19.30%, 9.92%, 9.26%, 3.03% and 2.87% respectively. In the test group, Candida was the most dominant fungus, followed by Aspergillus, Penicillium, Diutina, Apiotrichum, and Rhodotorula, with proportions of 61.34%, 15.18%, 5.98%, 5.23%, 2.82%, and 2.47%, respectively. The relatively average abundance of the mycobiota at the genus level is shown in Fig. 3e. The composition of these two groups of intestinal flora at the species level was analyzed to further understand the species-specific effects of AP on the intestinal flora (Fig. 3f).

Fig. 3

Flora composition and comparison of fungal microbiota composition in both groups. (a-f) Microbiota composition at the (a) phylum, (b) class, (c) order, (d) family, (e) genus and (f) species level

The clustered heatmap of the highest abundance features (n = 20) shows similar results. At the phyla level, the abundance of Basidiomycota, Rozellomycota and Chytridiomycota was found higher in the control group (Fig. 4a). Figure S2 shows the similarity or difference relationship between the community structure of the disease group and the healthy group by Community heatmap analysis at the class, order, and family level. At the genus level, the abundance of Ganoderma, Aspergillus, Auricularia, Talaromyces, Exophiala and Trichoderma was detected higher in healthy participants, while the high abundance of Candida, Diutina and Rhodotorula were detected in AP patients. At the species level, the abundance of Ganoderma sichuanense, unclassified Monascus, Talaromyces rugulosus, unclassified Auricularia, Penicillium oxalicum and Aspergillus amstelodami was higher in the control group while a higher abundance of unclassified Candida, Penicillium bialowiezense, Candida parapsilosis, Diutina catenulata, Rhodotorula mucilaginosa and Aspergillus fumigatus was found in the test group.

Fig. 4

Group clustering heat map of top 20 features in different taxa. (a) Phylum, (b) Genus, (c) Species

The Circos graph effectively showed the link between species and samples at the phylum, genus, and species levels, helping understand the diseased and healthy groups’ species distributions (Figure S3). The heatmap analysis on the OTU level of fungal microbiota showed a distinct microbiome difference between the two groups. There were 8 OTUs that were determined to be distinct between the sample groups. Out of these OTUs, one was found to be more prevalent in the test group compared to the control group. This particular OTU belongs to the genus Candida. The test group exhibited lower levels of 7 genera, including Aspergillus, Penicillium, Auricularia, Cladosporium, and unclassified Eurotiomycetes (Figure S4).

Significant shifts in the fungal microbiota between the groups

The Wilcoxon rank-sum test revealed significant differences in the fungal composition between the groups. Ascomycota were substantially more abundant in the test group than in the control group at the phylum level. In contrast, the abundance of Basidiomycota was significantly reduced in the test group compared to the control group (Fig. 5a). The test group exhibited a significantly higher abundance of Candida at the genus level compared to the control group. Conversely, the levels of Penicillium, Auricularia, unclassified Eurotiomycetes, Epicoccum and Vishniacozyma were significantly lower in the test group compared to the control group (Fig. 5b). Results obtained at the species level exhibit a resemblance to the outcomes obtained at the genus level (Fig. 5c).

Fig. 5

Flora composition and comparison of gut mycobiota composition in both groups. (a) Differences in fungi composition at the phylum level. (b) Differences in fungi composition at the genus level. (c) Differences in fungi composition at the species level. (d) Mycobiota differences at the genus level, as assessed by Linear discriminant analysis effect size

To establish the unique fungal taxa and prominent fungus in AP patients, a LEfSe was utilized to reveal the maximum difference in fungal microbiota structures between the groups. This analysis showed that the abundance of Candida was significantly higher in the test group than in the control group. The LEfSe analysis drew conclusions similar to those mentioned previously (Fig. 5d).

To evaluate the abundance network of gut fungal communities in both test and control groups, the top 50 species in terms of total abundance at the genus level were chosen and analyzed using Spearman’s correlation analysis to reveal the relationships between species. The size of the graph nodes represents species abundance, while different colors signify species. Positive and negative correlations are indicated by red and green connecting lines, respectively. The correlation coefficient increases with line thickness, and the more lines, the closer the species are to other species. Our analysis revealed a higher frequency and stronger correlation of gene interactions within the test group compared to the control group. In the test group, fungal genera were largely positively linked compared to the control group (Fig. 6a and b).

Fig. 6

Network analysis in the control and test group. (a) The distinctive gut fungal co-abundance networks in the control group. (b) The distinctive gut fungal co-abundance networks in the test group. (c) GMHI between the control group and the test group. (d) MDI between the control group and the test group. (e) The correlation analysis between GMHI and Simpson diversity. (f) The correlation analysis between GMHI and Shannon diversity

Using the gut microbiome health index (GMHI) and microbial dysbiosis index (MDI), we found that the GMHI was considerably greater in the healthy group compared to the diseased group (Fig. 6c). Additionally, the healthy group had a much lower MDI than the afflicted group (Fig. 6d). These findings suggest that the presence of fungal species linked to good health in the gut of the control group may be more abundant. In addition, we discovered that the Shannon and Simpson indices had a positive correlation with the GMHI, despite the fact that this result did not indicate a statistically significant difference (Fig. 6e and f).

Association of gut mycobiota with clinical and laboratory indicators

The study employed Spearman correlation analysis to evaluate the association between the dominating fungus of the top 50 and clinical variables. Our findings indicate that all genera, with the exception of Candida and unclassified Eurotiales, showed a positive correlation with CTSI. A generally negative association was found between PCT and fungus, as well as between D-dimer and fungus dominant microorganisms. Among them, Aspergillus showed a notable positive association with WBC levels, while unclassified Rozellomycota had a strong positive correlation with IL-6 (Fig. 7).

Fig. 7

Spearman correlation analysis between the fungus with clinical and laboratory indicators

This finding was further validated by linear regression analysis, which involved examining the relationship between clinical variables and the proportion of different fungal species in the community (Fig. 8a and e). Furthermore, the presence of Saccharomycetales showed a positive correlation with the WBC, in addition to the Aspergillus (Fig. 8b). And IL-6 had a favorable correlation with Diutina, Dirkmeia, and Acremonium, in addition to the unclassified Rozellomycota (Fig. 8f-h). However, the unclassified Hypocreales and Cystobasidium showed a negative correlation with CRP (Fig. 8c and d). Through the utilization of ordinal regression analysis, a statistical examination was conducted to assess the relationships between alpha diversity and clinical variables, as well as between beta diversity and clinical variables. The investigation revealed that none of the obtained results exhibited a significant difference. The Shannon index had a positive correlation with CTSI and WBC and a negative correlation with PCT (Figure S5a-c). Additionally, PCoA had a positive correlation with IL-6 and WBC, while it displayed a negative correlation with PCT (Figure S5d-f).

Fig. 8

Multivariate Association with Linear Models analysis between the clinical characteristics and species. (a) Between Aspergillus and WBC. (b) Between unclassified Saccharomycetales and WBC. (c) Between unclassified Hypocreales and CRP. (d) Between Cystobasidium and CRP. (e) Between unclassified Rozellomycota and IL-6. (f) Between Dirkmeia and IL-6. (g) Between Diutina and IL-6. (h) Between Acremonium and IL-6