2.1 Fatty acid composition of lipid samples

The fatty acid composition of the studied lipids was shown in Table 1. The most abundant fatty acids in SBO were C16:0, C16:1, C18:1 and C18:2, accounting for more than 90% of the total fatty acids. The top three fatty acids in FO were C20:5, C22:6 and C18:2, respectively, and the content of EPA and DHA in the total fatty acids exceeds 65%. The highest content of fatty acids in SL were EPA and DHA, and the content of other fatty acids were also relatively rich, among which the content of EPA/DHA accounts for about 44.8% of the total fatty acids.

Table 1 Fatty acid composition2.2 Serum lipid profiles of SD rats after 6-week of different dietary lipid intake

After six weeks of treatment with different dietary lipids, there were significant differences in blood lipid levels in SD rats (Fig. 1). The serum low-density lipoprotein cholesterol (LDL-C) (Fig. 1A) and high-density lipoprotein cholesterol (HDL-C) (Fig. 1B) concentrations in Ctrl group were significantly higher than those in SL, FO, and SBO groups (p < 0.01). The HDL-C concentration in SBO group was higher than that in FO and SL groups. The concentration of serum triglycerides (TG) (Fig. 1C) in SL group was significantly higher than that in Ctrl, SBO and FO groups (p < 0.01). The TG concentration in Ctrl and FO groups was significantly different from that in SBO group (p < 0.05). The serum total cholesterol (TC) concentration in FO group (Fig. 1D) was significantly higher than that in Ctrl, SL and SBO groups (p < 0.001).

Fig. 1

The serum lipid profifiles of the SD rats after 6-week lipid oral gavage administration. A Low-Density Lipoprotein Cholesterol; B High-Density Lipoprotein Cholesterol; C Triglycerides (TG); D Total Cholesterol (TC). The significance level was set at 0.05, ***p < 0.001, **p < 0.01, *p < 0.05

2.3 SCFA concentration in colon contents

The concent of total SCFAs in the Ctrl group was higher than that in the different oil feeding groups (Fig. 2). Except for the Ctrl group, acetic acid and propionic acid were the highest in FO group, followed by SL group, and the lowest in SBO feeding group. The concent of butyric acid was not statistically significant between the Ctrl and SBO groups (p > 0.05). The isobutyric acid concent of SBO was significantly lower than that of SL and FO groups (p < 0.01), and there was no difference between SL and FO groups (p > 0.05). There was no significant difference in valeric acid concent among Ctrl and SBO (p > 0.05). The content of isovalerate in Ctrl group was significantly different from that in SL, FO and SBO groups (p < 0.001), there was no significant difference in isovaleric acid concent among SL, FO and SBO (p > 0.05).

Fig. 2

SCFAs concent. A Acetic acid; B Propionic acid; C Butyric acid; D Isobutyric acid; E Valeric acid; F Isovaleric acid. The significance level was set at 0.05, ***p < 0.001, **p < 0.01, *p < 0.05

2.4 Cecal microbiota diversity2.4.1 Alpha diversity index analysis

The Observed species and Chao1 indices reflect the abundance of species, while the Shannon index and Simpson index reflect species diversity. As depicted in Fig. 3, the intestinal microbiota communities of FO and SBO groups exhibit higher values for both Observed species (Fig. 3A) and Chao1 (Fig. 3B) indices compared to the Ctrl group. Moreover, there were no significant differences in microbial diversity between SBO and FO groups, FO and SL groups. The Observed species and Chao1 indicated that the intestinal microbial communities in SBO and FO groups were significantly differ from those in the Ctrl group (p < 0.01). Furthermore, although there were no significant differences observed in the Shannon index (Fig. 3C) or Simpson index (Fig. 3D) among FO, SBO, and SL groups (p > 0.05), it is noteworthy that the Shannon index was highest in the SBO group followed by SL and FO groups, respectively, whereas the Simpson index was lowest in the SBO group followed by FO and SL.

Fig. 3

Alpha diversity index for each group. A Observed species; B Chao 1 index; C Shannon index; D Simpson index. The significance level was set at 0.05, ***p < 0.001, **p < 0.01, *p < 0.05

2.4.2 Beta diversity analysis2.4.2.1 Principal coordinate analysis (PCoA)

The weighted Unifrac algorithm was used in this study to comprehensively consider both species presence and abundance. Abundance was then weighted to calculate the distance between samples, allowing for observation of population differences between different samples or populations. As shown in Fig. 4A, the first principal component contributed 14.67% to the difference between samples, while the second principal component contributed 11.72%. The results showed that different dietary lipids altered the intestinal microbiota of SD rats, which was significantly different from that of the Ctrl group. The intestinal microbiota of the groups treated with different dietary lipids and the Ctrl group were distinctively separated. Moreover, in terms of spatial distribution, the SL processing group exhibited closer similarity to the Ctrl group.

Fig. 4

A Weighted Unifrac algorithm PCoA analysis; B Weighted Unifrac algorithm NMDS analysis

2.4.2.2 Nonmetric multidimensional scaling (NMDS) analysis

The results of weighted Unifrac nonmetric scaling show that each point in the figure represents a sample, while different colors of the points were used to represent groups of samples. The closer the distribution of points, the more similar the samples were represented. There was significant clustering of microbial communities among different groups. The results indicated significant differences between the Ctrl group and the FO, SBO and SL treated bacterial communities (Fig. 4B).

2.4.2.3 Heatmap analysis

A heatmap was generated using a distance algorithm and the R language tool (Additional file 1: Fig. S1). The similar samples were clustered closer together in the tree, which reflectd the similarity and difference in species composition at the genus level for all samples. The color represented the abundance of species, with a gradient from blue to red indicating increasing abundance.

2.5 Microbiota composition of cecum2.5.1 Venn diagram analysis of species

In order to investigate the regulatory effects of different dietary fats on intestinal microbiota, we employed high-throughput sequencing of bacterial 16S rRNA to determine the composition of the bacterial community in the sample. Therefore, we sequenced the original cecal microbiota of rats fed different dietary lipids for 6 weeks. Based on the analysis of Fig. 5, a total of 412 shared operational taxa (OTUs) were observed across all four groups, indicating a presence of at least 412 species of symbiotic bacteria in these groups. The number of specific OTUs in Ctrl, FO, SBO and SL groups were found to be 404, 395, 672 and 452 respectively, suggesting distinct microbial species associated with each component.

Fig. 5

Venn diagrams of OUT distribution in all 4 groups

2.5.2 Analysis of community variability between groups2.5.2.1 Phylum level

The dominant bacteria identified at the phylum level in the collected door sample were Firmicutes, Actinomyces, Proteobacteria, Desulphuricobacteria, Bacteroidetes, Patescibacteria and Campylobacter (Fig. 6A). Among them, Firmicutes, Actinomycetes and Proteobacteria were found to be the most abundant. The average relative abundance of Firmicutes was highest in the Ctrl group (87.94%), followed by Proteobacteria (5.45%) and Actinobacteria (3.54%). In the FO group, the average relative abundance of Firmicutes (84.37%), Proteobacteria (6.12%) and Actinobacteria (3.36%) was highest. The top three bacteria in SBO group and SL group were Firmicutes, Actinobacteria and Desulphurobacteria respectively. Among them, Bacteroidota had the highest number in Ctrl group (87.94%), while it had the lowest number in SL group (3.03%). As shown in Fig. 6, there was a similar dominance of bacteria at the gate level among Ctrl, FO, SBO and SL groups. Interestingly, we found that Firmicutesto count was highest in FO group whereas significantly lower than Ctrl for SBO and SL groups (Fig. 6B). Compared to Ctrl group Bacteroidetes to count increased for FO, SBO and SL groups, the ratio offirmicutes to Bacteroidetes (F/B) was 126.97 for Ctrl group which was significantly higher than that for SL group (the lowest being 28.30). Abundance increased significantly for FO, SBO and SL groups compared to Ctrl.

Fig. 6

Histogram of structure at the level of Phylum and Gennus. A Relative abundance; B Absolute abundance; C Relative abundance; D Absolute abundance

2.5.2.2 Genus level

At the genus level (Fig. 6C), there were no significant differences in the proportion of major flora among different dietary groups compared to the control group. The proportion of major flora mainly consisted of Romboutsia, Lactobacillus, Unassigned, Turicibacter, Clostridium_sensu_stricto_1 and Ligilactobacillus. Through absolute quantification (Fig. 6D), we found that the FO group showed an increased abundance of Romboutsia, Lactobacillus, Ligilactobacillus, Escherichia-Shigella and Lachnospiraceae_NK4A136_group compared to the Ctrl group. Specifically, Desulfovibrio and NK4A214_group significantly increased in abundance while Klebsiella and Blautia both decreased significantly in the FO group. Among them, Blautia exhibited the most pronounced decline in the SL group. Additionally, it is worth noting that with regard to Blautia in SL group declining the most.

2.6 Categorical biomarker analysis

To further investigate the differences in intestinal microbiota among the Ctrl, FO, SBO, and SL groups, we employed the linear discriminant analysis effect size (LEFSe) (Fig. 7) and Linear Discriminant Analysis (LDA > 6) (Fig. 8) methods to identify specific taxa that were differentiated at various classification levels within each group. In the Ctrl group, two dominant bacterial groups were observed, Lactobacillus_intestinalis and Lachnospiraceae_bacterium_19gly4. The FO group exhibited ten dominant bacterial groups, including Helicobacter, Campylobacterota, and Helicobacteraceae, Campylobacteria with Campylobacterales showing the highest score. Similarly, the SBO group displayed ten dominant bacteria such as Dubosiella and Dubosiela.s_uncultured_bacterium which also scored high on LDA. Lastly, in the SL group, six dominant bacterial groups were identified including Butyricimonas and Turicibacter.s_uncultured_bacterium which also scored high on LDA. These findings indicated that specific bacterial taxa identified at different classification levels varied among the treatment groups.

Fig. 7

LEfSe branching diagram. Different colors repr; sent different groups. For instance, red nodes in branches indicate species with significantly higher abundance in the red group, green nodes indicate species with significantly higher abundance in the green group, and yellow nodes represent species with no significant difference in abundance between groups. The diameter of each node is proportional to its abundance. The nodes of each layer depict phyla, class, order, family, genus, species from the inside out, while the annotations on each layer indicate phyla, class, order, family, genus, species from the outside in. Annotations at the gate level are displayed on the outermost ring, whereas annotation information at other levels is provided in the legend on the right

Fig. 8

LDA score chart. The bars of various colors represent different species with an LDA score (log10) greater than 2 in distinct groups and significantly high abundance within each group. The length of the bar indicates the magnitude of the LDA score value