Diet
Deviations in the relative abundance of DNA sequences were used to conclude differences in prey ingestion among different seasons. Total dissimilarities in prey consumption were inferred by comparing matrices of pairwise proportion among prey taxon present in each population. Across three seasons, analysis of chipmunk’s feces revealed 27 phyla representing up to 33 different genera in the chipmunk’s diet. Dietary composition varied seasonally (P < 0.001, Fig.2, Fig. 2A, Additional file 1 and Additional file 4). Ascomycota (43.8% of sequences on average), Streptophyta (37.0%), Nematoda (15.8%), Basidiomycota (2.1%), Arthropoda (0.3%), Glomeromycota (0.2%) and Zygomycota (0.1%) are the dominant phyla that was consumed by chipmunk among three seasons. The most abundant phylum Ascomycota including genus Aspergillus, Penicillium, Aureobasidium and phylum Streptophyta consist of genus Enemion and Hylomecon were observed in spring. Maximum abundance of phylum Basidiomycota including genus Hymenogaster and phylum Nematoda consist of genus Travassostrongylus, Oswaldocruzia, Murshidia, and Steinernema was observed in the autumn diet (Additional file 1).
We employed the Chao1 estimator of total richness to estimate the number of phylum and genus present in the samples (Fig. 2A, C). The Shannon index (H) that correlates positively with species richness and evenness was also calculated at both phylum and genus levels (Fig 2B, D). Overall, Chao1 and Shannon diversity indices indicated the great richness in diet items. We observed maximum α-diversity in the autumn season (Fig.2). whereas, spring season had significantly higher variance in their α-diversity compared to summer and autumn seasons. The diet during the summer and autumn season was similar but variation was observed in the spring season. In contrast to the analysis at the phylum level and Genus level, revealed that the diet richness and biodiversity were maximum in autumn.
A SIMPER and ANOSIM test recognized significant differences, indicating an overall variation in relative proportions of prey items. Similarity percentages (SIMPER) exploration reveal the amount of input of each taxon based on dissimilarity observed between groups. It allowed us to identify which phylum and genus were most significant in generating the observed pattern of dissimilarity. All samples were pooled to perform one overall multi-group SIMPER and consumed the Bray-Curtis measure of dissimilarity, relating in turn, the individual sample in spring with the individual sample in summer and autumn. SIMPER also let us to identifying the taxon that was likely to be the dominant contributors to any difference between assessed seasons (Additional file 3). At the phylum level, the overall observed average dissimilarity in diet among seasons was 21.87% and the overall average Bray-Curtis dissimilarity at genus level among seasons was 69.94%. At the phylum level, the Nematoda (24.09%), Ascomycota (17.43), Basidiomycota (11.02%) and Apicomplexa (3.27%) contributed most in the differences between seasons diets (Additional file 3). Results from ANOSIM showed more dissimilarity at genus level as compared to phylum level. Diet composition at phylum level showed a significant difference between summer and autumn season (SIMPER, Euclidian distance, p<0.05, p=0.04, R = 0.1949, Permutation N: 9999) but illustrated highly significant dissimilarity between Autumn and spring season (SIMPER, Euclidian distance, p=0.001, Permutation N: 9999). Genus level diet composition is highly significant (p<0.001, R = 0.2354) between spring/summer and spring/autumn.
Principal component analysis (PCA) was used to emphasize variation, bring out strong patterns of correlation among seasons. Bi-plot was used to assess the data structure. Results revealed that spring has more dissimilarity from other two seasons and summer showed more similarity with autumn. Among diet items Travassostrongylus, Aspergillus, Penicillium, and Enemion are the key genus cause dissimilarity among seasons, PC1 shows 69.78 % variances whereas PC2 shows 16.33% variance (Fig. 3A).
Microbiota
Chipmunk’s gut microbiome unveiled 10 phyla including up to 15 different genera. Gut microbiome composition varied seasonally, and a summary of all the comparisons made in this way is given in Fig.2, Fig. 4 and Additional file 2 (P < 0.001). The four most abundant microbial phyla were Firmicutes (55.2%), Bacteroidetes (30.4%) and Proteobacteria (6.8%). Firmicutes consist of Lactobacillus genus, Bacteroidetes comprised of Prevotella genus and Proteobacteria contained Desulfovibrio genus were the more abundant phylum in the spring. These phyla exhibited a significant temporal change in relative abundance over the three seasons, decreasing more than two fold from the gut microbial community in spring. We observed that the a-diversity of the gut microbiome of chipmunk varied seasonally (P < 0.001, Additional file 3 and Additional file 2).
A preliminary Principle Component Analysis (PCA) was conducted to visualize differences in bacterial genus composition between seasons, and to determine which genus were most strongly associated with the differences observed. PCA confirmed that samples from the spring season formed a distinct position in the ordination plot and summer showed similarity with autumn. This separation was most apparent along the PC1 axis which explained 93.62% of the overall variation, and for which Lactobacillus had maximum dissimilarity among seasons and was abundant in spring (Fig.4). The PC2 axis explained only 6.33% of the overall variation; however, no distinctions between seasonal groups were made through this component.
Differences in ß-diversity (i.e., SIMPER and ANOSIM) were also measured. When all time points were averaged together, there was significant variation among the three seasons. SIMPER analyses across phylum and genus levels were employed to identify taxa with the highest contribution to differences between the diet types. Further, SIMPER analyses across phylum and genus level were employed to identify taxa with the highest contribution to differences between the seasons. (Additional file 3). The distinction between diets types was more apparent as the taxonomic level became more specific where SIMPER detected 34.37% dissimilarity at the phylum level, which increased to 35.22% dissimilarity at the genus level. At the phylum level, Proteobacteria (35.45%), Firmicutes (26.7%) and Bacteroidetes (23.59%) were shown to be better discriminators as they contributed most to differences between seasons, it indicates that the gut microbiome undergoes dramatic seasonal fluctuations. When we compared three seasons, significant (SIMPER, Euclidian distance, p<0.05, p=0.015, R=0.0730, Permutation N: 9999) dissimilarity was observed between autumn and spring season at the phylum level. Like diet composition, microbiome showed highly significant dissimilarity (p<0.001, R=0.235) between spring/autumn and spring/summer seasons at the genus level. Gut microbiome showed no statistical difference (SIMPER, Euclidian distance, p>0.05, p=0.59, Permutation N: 9999) between summer and autumn season.
Predicted gut microflora function using PICRUSt
It is unclear that seasonal variation in the gut microbiome affects host metabolism. To understand the specific effects of the gut microbiome on host metabolism, PICRUSt was performed to predict the chipmunk gut microbiome functions, which showed that chipmunk in different seasons exhibited some differences in metabolism abundance at level 2, including carbohydrate, protein, amino acid, Xenobiotics, energy, Cofactors and Vitamins, Glycan, Lipid, Terpenoids, and Polyketides metabolism. A Violin Plot was used to visualize the distribution of the data and its probability density (Fig. 5). The abundance of carbohydrate, Lipid, Nucleotide, Xenobiotics, energy, Terpenoids and Polyketides metabolism were higher in spring than in summer and autumn.
Moreover, spring and summer showed significant differences (ANOSIM, P < 0.001, p= 0.005, Euclidian R= 0.20), whereas a highly significant difference was observed between spring and autumn (ANOSIM, p=0.0005) and insignificant Euclidian difference (p=0.29) was investigated while comparing summer with autumn. SIMPER analyses among three seasons observed average Bray-Curtis dissimilarity was 3.627% as shown in Additional file 4. This analysis assisted us in figuring out the relationship between seasonal metabolic capacities.