Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/springer-journals/effects-of-dietary-supplementation-with-prebiotics-and-pediococcus-9omaZhAbBn
- Publisher
- Springer Journals
- Copyright
- Copyright © The Author(s) 2023
- eISSN
- 2524-4671
- DOI
- 10.1186/s42523-023-00228-w
- Publisher site
- See Article on Publisher Site
Abstract
Background Given the importance of gut microbiota for health, growth and performance of the host, the aquacul‑ ture industry has taken measures to develop functional fish feeds aiming at modulating gut microbiota and inducing the anticipated beneficial effects. However, present understanding of the impact of such functional feeds on the fish is limited. The study reported herein was conducted to gain knowledge on performance and gut health character‑ istics in post‑smolt Atlantic salmon fed diets varying in content of functional ingredients. Three experimental diets, a diet containing fructo‑ oligosaccharides (FOS), a diet with a combination of FOS and Pediococcus acidilactici (BC) and a diet containing galacto‑ oligosaccharides (GOS) and BC, were used in a 10‑ weeks feeding trial. A commercial diet with‑ out functional ingredients was also included as a control/reference. Samples of blood plasma, mucosa and digesta were subjected to microbiota, transcriptome and metabolome profiling for evaluation of the diet effects. Results No significant growth differences were observed between fish fed the supplemented diets, but FOS–BC fed fish showed significantly faster growth than the control fed fish. The microbiota results showed that the BC was present in both the digesta, and the mucosa samples of fish fed the FOS–BC and GOS–BC diets. Digesta‑associated microbiota was altered, while mucosa‑associated microbiota was relatively unaffected by diet. Replacing FOS with GOS increased the level of metabolites linked to phospholipid, fatty acid, carnitine and sphingolipid metabolism. Variation in metabolite levels between the treatments closely correlated with genera mainly belonging to Firmicutes and Actinobacteria phyla. The transcriptome analyses indicated diet effects of exchanging FOS with GOS on immune functions, oxidative defense and stress responses. No significant diet effect was observed on intestinal inflammation in the pyloric caeca or in the distal intestine, or on lipid accumulation in the pyloric caeca. Conclusions Dietary supplementation with BC induced moderate effects on the microbiota of the digesta, while the effects of replacing FOS with GOS were more marked and was observed also for nutrient metabolism. Our data indicates therefore that the quality of a prebiotic may be of great importance for the effects of a probiotic on gut microbiota, function, and health. *Correspondence: Anusha K. S. Dhanasiri anusha.dhanasiri@nmbu.no Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 2 of 22 Keywords Functional ingredients, Prebiotics, Probiotics, FOS, GOS, Gut microbiota, Pediococcus acidilactici, Metabolomics, Transcriptomics, Atlantic salmon Background species have increased over the past years including To be able to grow sustainably, the salmon aquaculture some studies on salmonids [2, 5, 15, 16]. Dietary appli- industry has during the last 2 decades moved away from cation of P. acidilactici and GOS has shown effects such the traditional high fishmeal/fish oil diets, by gradually as increased immune responses and disease resistance, increasing the use of plant raw materials and alterna- microbiota and metabolic alterations in rainbow trout tive sources of lipid. Dietary incorporation of functional (Oncorhynchus mykiss) [17–19], increased growth in ingredients is also gaining attention to improve the juvenile rockfish (Sebastes schlegeli) [6] and some effects robustness of the fish. Gut microbiota is important for on mucosal and serum immune parameters in common performance and well-being of the fish. Therefore, efforts carp (Cyprinus carpio) fingerlings [20]. A few studies have been made to develop functional feeds aiming at have reported effects of dietary inclusion with P. acidilac - modulating the gut microbiota to induce anticipated tici and FOS such as modulation of gut microbiota and beneficial effects. Several previous studies have been con - immunity in Atlantic salmon [7] and increased growth ducted to evaluate the effect of feeds supplemented with performance of Caspian roach (Rutilus frisii kutum) fry probiotics, prebiotics or synbiotics, i.e. combinations of [21]. pre and pro-biotics, for farmed fish species including Economically, Atlantic salmon is one of the most Atlantic salmon [1–5]. However, further efforts are still important farmed fish species worldwide [22]. The post- needed to better understand the combined effect of those smolt stage (early marine phase) is one of the critical functional ingredients on gut microbiota, gut function stages in Atlantic salmon life cycle [23]. Suppression of and health, and overall performance of the fish. gut health [12] and alterations of gut microbiota [24] Dietary supplementation of probiotic bacteria can were reported in Atlantic salmon during early marine modulate gut microbiota and gut immune responses in phase. In this stage functional feeds could play an impor- beneficial ways and contribute to the synthesis of nutri - tant role in increasing survival, health, growth, and ents, ultimately improving disease resistance and growth overall performance of the fish. Considering the impor - performance of the fish [1]. The lactic acid bacteria, tance of gut microbiota in modulating the gut health Pediococcus acidilactici MA 18/5M, is among the most and ultimately overall health and performance of the widely studied probiotic bacteria for farmed fish species fish, this study was conducted to evaluate effects in post- [6–10] and has been reported to enhance gut mucosal smolt Atlantic salmon of supplementing a diet contain- and peripheral immunity. Prebiotics may also exert ben- ing the prebiotic FOS with the probiotic P. acidilactici eficial host effects, via stimulation of the growth and/ (BC) and replacing FOS in the diet containing BC with or the activity of the gut microbial population [3]. Sev- GOS. A grower diet without functional ingredients was eral studies have indicated beneficial effects in fish of also included as a reference/control. An overview of the prebiotics such as fructo-oligosaccharide (FOS), galacto- experimental set up and investigated endpoints is illus- oligosaccharide (GOS), mannan-oligosaccharide, beta trated in Fig. 1 and detailed in the materials and methods glucans and inulin [3, 4, 11]. On the other hand, a recent section. A multi-omics analytical approach was chosen large-scale study with salmon under commercial farm- with microbiota, transcriptome and metabolome profil - ing conditions showed little or no effects of dietary sup - ing. This study strengthens the knowledge basis of effects plementation of a mixture of nucleotides, yeast cell walls of use of functional feeds on fish by unveiling the com - and essential fatty acids [12], but indicated that these plex interrelated associations among the gut microbiota– specific functional ingredients may represent an ener - transcriptome–metabolites. The knowledge gain would getic cost for the fish. also aid in optimizing the inclusion of functional diets Synbiotics, a mixture of probiotic and prebiotic agents, into commercial feed formulations. can have beneficial effects on the host by improving the survival and implantation of probiotic and/or the growth Results and activity of the indigenous beneficial bacteria in the Detailed comparisons are made between the two pairs of gut [13]. Therefore, an optimal combination of probiotics treatment for which the cause of differences can be inter - and prebiotics in a single product could elicit a superior preted and discussed to achieve the goals of the study, effect, compared to the activity of each component alone i.e. fish fed the FOS and the FOS–BC diets and those fed [14]. Studies of application of synbiotics in aquaculture the FOS–BC and GOS–BC diets. This approach will help Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 3 of 22 Study parameters Analytical process Performance Growth analysis Feeding trials Expected outcome Histological Gut health Reference analysis FOS diet Diet 16S rRNA Gut microbiota sequencing FOS-BC diet Transcriptomics Gene expression RNA-seq GOS-BC diet Metabolomics Metabolic effect UPLC-MS/MS SCFAs profiling SCFAs changes LC-MS/MS Fig. 1 Schematic representation of experimental design. This study evaluated the effects of supplementation of P. acidilactici in diets for Atlantic salmon performance and gut health after transfer from freshwater to seawater. Fish were fed FOS alone (FOS diet) and FOS and GOS in combination with P. acidilactici (FOS–BC and GOS–BC diets respectively) and a commercial diet as a control/reference for 10 weeks. Six different parameters were analyzed using traditional and state‑ of‑art ‑multi‑ omics techniques as detailed in the materials and methods section to investigate the effects of supplementing the diet containing the prebiotic FOS with the probiotic P. acidilactici (FOS–BSC vs. FOS) and replacing FOS in the FOS–BC diet with GOS (GOS–BC vs. FOS–BC) on post‑smolt Atlantic salmon. Photograph. Geir Mogen, BioMar us to understand the effects of supplementation of pro - biotic, BC, to a diet containing prebiotic, FOS, and the influence of alteration of prebiotic combined with BC. Performance data 3.5 ab ab The fish grew well throughout the experiment showing 3.0 thermal growth coefficients (TGCs) averaging about 3.1 2.5 (Fig. 2). Fish in the FOS–BC group grew significantly faster than those in the control/reference group, showing 2.0 TGCs of 3.23 and 2.96, respectively, during the 10 weeks 1.5 of feeding. However, no significant differences in growth were observed with the supplementation of BC to FOS 1.0 diet (FOS–BC vs. FOS) or after replacing FOS with GOS 0.5 in FOS–BC diet (GOS–BC vs. FOS–BC). Feed intake 0.0 and feed conversion ratios, which averaged 847 g ± 8 ControlFOS FOS-BC GOS-BC (SEM) and 1.12 ± 0.02 (SEM), respectively, showed no significant differences among the four treatments. Feed groups Fig. 2 The thermal growth coefficient ( TGC) of Atlantic salmon fed different diets. Post ‑smolt Atlantic salmon was fed a commercial diet Gut histology as a control/reference and three experimental diets: FOS alone (FOS The distal intestine and pyloric caeca of the fish from the diet) and FOS and GOS in combination with P. acidilactici (FOS–BC four treatments showed largely normal morphological and GOS–BC diets respectively) for 10 weeks. Values are mean of characteristics, but some individuals from all diet groups 210 fish per group. Error bars represent SEM (standard error of the showed abnormal morphology that ranged from mild to mean). Different letters among values indicate statistically significant differences (q ≤ 0.05). Values sharing the same letters are not severe. Figure 3a and b illustrate the observations made statistically significant. Significant difference observed only between regarding signs of inflammation in the distal intestine, the fish fed Control and FOS–BC diets (q ≤ 0.05) i.e. regarding cell infiltration and loss of distal intestine Thermal growth coefficient (TGC) Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 4 of 22 (a) DI mucosal cell infiltration (b) Loss in DI enterocyte vacuolization 16 16 12 12 8 8 4 4 0 0 ControlFOS FOS-BC GOS-BC ControlFOS FOS-BC GOS-BC (c) PC mucosal cell inflitration (d) PC enterocyte steatosis 16 16 12 12 8 8 4 4 0 0 ControlFOS FOS-BC GOS-BC ControlFOS FOS-BC GOS-BC Normal Mild Moderate Marked Severe Normal Mild Moderate Marked Severe Fig. 3 Histomorphological evaluation of distal intestine (DI) and pyloric caeca (PC) of Atlantic salmon. Number of fish scored as normal, mild moderate, marked, or severe for selected histomorphological of a distal intestine inflammatory cell infiltration (p = 0.638), b loss of distal intestine enterocyte vacuoles (p = 0.097), c inflammatory cell infiltration of the pyloric caeca mucosa (p = 0.529), and d lipid accumulation (steatosis) in pyloric caeca enterocytes (p = 0.437). p values represent outcomes of an ordinal logistic regression for differences in histology score outcomes between the treatment and the reference group, control enterocyte vacuoles, respectively. The results showed diet and FOS–BC diet (observed ASVs: p = 0.02 and no significant differences between treatments. The same Shannon: p = 0.005). The mucosa samples did not show was observed regarding infiltration of inflammatory cells significant diet effects among the fish fed different diets. in mucosa and lipid accumulation (steatosis) in pyloric caeca, (i.e. inflammation and steatosis, Fig. 3c and d, Beta diversity respectively). The gut histological parameters were not Beta diversity, i.e. differences in bacterial taxa between affected by either supplementation of BC to FOS diet or samples, taking into account taxa differences as well after replacing FOS with GOS in the FOS–BC diet. as the abundance of the taxa, was evaluated by PER- MANOVA analysis based on Bray–Curtis dissimilarity Gut microbiota profiling matrix at ASV level. For the digesta samples (Fig. 4a), The absolute bacterial DNA levels overall significant differences among treatments were Bacterial DNA levels measured by qPCR analysis did not observed (p = 0.03). The microbiota structure in fish from show significant differences between any of the three the FOS–BC treatment showed clear separation from experimental diets. However, the variation between sam- those in the FOS treatment (p = 0.007). On the other ples within treatment was large (Additional File 1: Fig. hand, the microbiota in fish from the GOS–BC treatment S1). Bacterial DNA levels in digesta were, in general, clustered close to, but distinct from that of the FOS–BC higher than the levels in mucosa. treatment (p = 0.02). The mucosa samples (Fig. 4b) did not show significant differences in beta diversity among different treatments. Alpha diversity Results regarding alpha diversity, i.e. number of different ASVs within a sample, measured as observed ASVs and Taxonomic composition Shannon indices, are presented in Additional File 1: Fig. In the digesta, at the phylum level, Firmicutes dominate S2a and b for digesta and S2c and S2d for mucosa. In the in most of the samples and Firmicutes and Proteobac- digesta samples, alpha diversity showed differing trend teria, represented more than 90% of the average rela- among the treatments (observed ASVs: p = 0.07 and tive abundance in all treatments (Additional File 1: Fig. Shannon: p = 0.08). However, pairwise comparisons indi- S3a). At the genus level, the lactic acid bacteria group, cated a significant difference between fish fed GOS–BC represented mainly by Lactobacillus and Leuconostoc No. fish assessed No. fish assessed No. fish assessed No. fish assessed Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 5 of 22 0.2 0.5 Feed groups 0.0 0.0 -0.5 -0.2 -1 -0.5 0.0 0.5 1.0 -0.3 00.3 0.6 NMDS1 NMDS1 Fig. 4 NMDS plots based on Bray–Curtis dissimilarity matrix showing beta diversity at ASV level. Beta diversity in the distal intestine digesta (a) and mucosa (b) of the Atlantic salmon fed with a control/reference diet and three experimental diets: FOS, FOS–BC and GOS–BC. The whole bacterial community of each sample is represented by a dot on the figure. Samples with similar bacterial compositions are closer to each other. PERMANOVA statistics for digesta: F value: 2.07; R‑squared: 0.18; p value: 0.03; and [NMDS] Stress = 0.13. PERMANOVA statistics for mucosa: F value: 0.66; R‑squared: 0.07; p value: 0.75; and [NMDS] Stress = 0.20. k value for NMDS analysis = 2 comprised around 50% of the average relative abundance for the digesta samples. In the digesta samples, the most in all treatments (Fig. 5a). The complete list of genera in important taxon which allowed discrimination of fish digesta which showed significant changes in their abun - fed diets supplemented with BC from the other fish, was dance among treatments are presented in Additional File P. acidilactici (Fig. 6a). In the mucosa, it was the fourth 2: Tables S1. The number of differentially abundant gen - most important discriminatory taxon (Fig. 6b). Both era in FOS–BC versus FOS comparison was 19, and 15 of digesta (Fig. 6c) and mucosa (Fig. 6d), samples from the them showed higher abundance in the FOS–BC fed fish. FOS–BC and GOS–BC diet fed fish had higher abun - Pediococcus and Staphylococcus were among the genera dance of P. acidilactici compared to the FOS and con- showing increase. Fish fed the GOS–BC diet, compared trol diet fed fish. Moreover, both digesta and mucosa to those fed the FOS–BC diet, showed reduction in 24 samples from the fish fed the GOS–BC diet had higher genera including Kurthia, Savagea, Staphylococcus, Vago- abundance of P. acidilactici compared to the fish fed the coccus and Peptostreptococcus. FOS–BC diet (Fig. 6c, d). The abundance of P. acidilactici In the mucosa, the most abundant phyla were Spiro- in digesta was substantially higher than its abundance in chaetes, Firmicutes and Proteobacteria. Together they mucosa samples. accounted for approximately 90% of averaged relative abundance in all the treatments (Additional File 1: Fig. Transcriptome profiling S3b). The dominant genera in mucosa were Brevinema, The RNA-seq data showed raw read counts ranging Aliivibrio and Lactobacillus which comprised around from 20.4 to 42.8 million reads with an average count of 70% averaged relative abundance per feeding group 30.1 million per sample. Uniquely mapped reads ranged (Fig. 5b). between 15 and 32 million among the samples having an We further employed the Random Forest model, a 71% of average unique mapping efficiency. supervised machine-learning algorithm, for classifica - tion and identification of microbial taxa that differenti - Differently expressed genes (DEGs) ate among treatments. Random Forest model performed The global transcriptomic analysis revealed the highest well for correctly predicting the microbial species of the number of DEGs (Benjamini–Hochberg adjusted p < 0.1, replicates fish from four treatments in the digesta sam - Table 1) in the GOS–BC treatment compared to the ples, but not in the mucosa samples, as indicated by 0.25 other treatments. Annotated DEGs among treatments are and 0.906 OOB (out of bag) error obtained, respectively presented in Additional File 3. Transcriptomic changes in (Additional File 2: Tables S2 and S3). Therefore, in the the distal intestine of fish fed the FOS–BC diet compared following, we mainly focus on digesta-associated micro- to those fed the FOS diet showed a low number of DEGs biota. The model classified the treatments FOS–BC and (27 up- and 6 down-regulated, Table 1, Additional File GOS–BC quite precisely with 87.5% predicting accuracy 3: File S1). Global transcriptome analysis showed major NMDS2 NMDS2 Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 6 of 22 FOS FOS-BC GOS-BC Control 1.00 1.00 0.75 0.75 0.50 0.50 0.25 0.25 0.00 0.00 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 Replicate fish from feed groups Weissella Lactobacillus Aliivibrio Kurthia Mycoplasma Leuconostoc Savagea Pediococcus Corynebacterium 1 Others Lactococcus FOS FOS-BC GOS-BC Control 1.00 1.00 0.75 0.75 0.50 0.50 0.25 0.25 0.00 0.00 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 Replicate fish from feed groups Brevinema Others Lactococcus Weissella Aliivibrio Leuconostoc Not-Assigned Kurthia Lactobacillus Mycoplasma Pediococcus Fig. 5 Top 10 most abundant genera of digesta (a) and mucosa (b) from distal intestine. The samples are grouped by feed groups: Atlantic salmon fed with a control/reference diet and three experimental diets: FOS, FOS–BC, and GOS–BC diets. The mean relative abundance of genera per feed group is presented on the right side differences in the distal intestine between fish fed the FOS–BC diet. Among the upregulated genes in fish fed GOS–BC diet and FOS–BC diet, 174 up- and 46 down- with GOS–BC diet were cysteine knot cytokine mem- regulated in fish fed GOS–BC diet compared to those fed bers, interleukin 17 and receptors, Il17a, il17a/f1 and Relative abundance Relative abundance Control Control FOS FOS FOS-BC FOS-BC GOS-BC GOS-BC Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 7 of 22 Digesta Mucosa Pediococcusacidilactici Lactobacillus fermentum Corynebacterium aurimucosum Lactobacillus manihotivorans Lactobacillus delbrueckii Leuconostoccitreum Lactobacillus salivarius Pediococcusacidilactici Staphylococcus sp. MRC5-3-1 uncultured-bacterium Ambiguous_taxa Lactobacillus sp LB12 uncultured Brevibacterium sp. Lactobacillus salivarius Pediococcusargentinicus uncultured Brevibacterium sp. Kurthiasp. PAOGL173 Microgenomates bacterium Lactobacillus sp. LA-6 Enterococcus cecorum c P. acidilactici in mucosa P. acidilactici in digesta d Feed groups Control FOS FO S-BC GOS-BC Control FOS FO S-BC GOS-BC Feed groups Feed groups Fig. 6 Random Forest importance plot indicating top 10 microbial species valuable for discriminating four treatments. Top 10 microbial species in digesta (a) and mucosa (b). The importance of the species is ordered from top to bottom and an estimate of their importance is indicated by the corresponding mean decrease accuracy. Color ranging from blue to red indicates the species abundance ranging from low to high i.e. blue color indicates low abundance and red color indicates high abundance. Box plots showing filtered absolute counts of P. acidilactici in digesta (c) and mucosa (d) which is important for separating fish in FOS–BC and GOS–BC from those in the control and FOS treatments. Note that the scale of y‑axis is different for digesta (c) and mucosa (d) in box plots Table 1 Number of differentially expressed genes (DEGs) and duox2) and NADPH oxidase activator 1 (noxo1a and resulted from pairwise comparisons of treatments noxo1b) and key antioxidant enzyme, glutathione peroxi- dase 1b (gpx1b). Comparisons Differentially expressed genes (DEGs) (q < 0.1, FC > 1.5) Total Upregulated Downregulated Gene ontology (GO) enrichment analysis FOS–BC versus FOS 34 27 6 Results of GO enrichment analysis did not indicate GOS–BC versus FOS–BC 220 174 46 enrichment of biological processes within the statistical FOS versus control 07 04 03 criteria for the FOS–BS versus FOS comparison due to FOS–BC versus control 07 02 05 the low number of DEGs. Statistically enriched biologi- GOS–BC versus control 537 269 268 cal processes, as indicated by upregulation of genes, were observed only for GOS–BC versus FOS–BC and GOS– BC versus Control. The complete list of summarized GO terms generated from respective comparisons are avail- i17ra; TNF superfamily members and receptors tnfrsf1b, able in Additional File 2: Table S4. The summarized GO tnfrsf1, tnfrsf9a and tnfsf18; beta trefoil cytokine fam- terms generated from enriched nonredundant biological ily member il-1rl; and a number of chemokines (Addi- function GO terms are presented in Fig. 7 for upregu- tional File 3: File S2). The fish in the GOS–BC treatment lated genes in fish fed the GOS–BC diet compared to also showed an increase in expression of transcripts of the FOS–BC diet. Among the enriched GO biological NADPH oxidases family of enzymes, dual oxidases (duox process terms were immune response, apoptotic process, Abundance Abundance Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 8 of 22 Fig. 7 Non‑redundant enriched gene ontology (GO) biological processes. Figure shows the enriched biological processes detected for the upregulated genes in Atlantic salmon fed the GOS–BC diet compared to fish fed the FOS–BC diet. Data are summarized as scatter plots using REVIGO tool. GO terms are marked with circles and plotted according to semantic similarities to other GO terms. The color of the circles ranging from yellow to red indicates the order of increase in log10 p value. Circle sizes are proportional to the respective frequencies of the GO terms (circles of more general terms are larger). Not all the terms are indicated in the figure due to the space limitations and the complete list of non‑redundant enriched GO terms can be found in Additional File 2: Table S4 inflammatory response, response to stress and reactive Table 2 Number of significantly altered metabolites obtained oxygen species metabolic process (Fig. 7). from pairwise comparisons of treatments Comparisons Significantly altered Significantly altered metabolites in digesta metabolites in plasma Metabolome profiling (p ≤ 0.05) (p ≤ 0.05) The global metabolome profiling detected 747 and Increased Decreased Increased Decreased 655 metabolites in total, respectively in distal intestine FOS–BC versus FOS 14 13 03 19 digesta, and blood plasma samples collected from the GOS–BC versus 86 23 18 34 various treatments. The number of significantly altered FOS–BC metabolites among fish fed different diets are presented FOS versus Control 60 56 104 48 in the Table 2. FOS–BC versus 63 63 65 60 All the detected metabolites highlighting the sig- Control nificantly altered metabolites in each of the com- GOS–BC versus 165 62 103 101 parisons between treatments are presented in the Control Additional File 3: Files S4 and S5 for digesta and Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 9 of 22 Short chain fatty acid levels plasma, respectively. Although some differences were The metabolome analyses of plasma samples did not observed, many of the changes in plasma and digesta show significant treatment effects, neither regarding metabolites mirrored each other by dietary treat- the major SCFAs (acetic acid, butyric acid, and propi- ment (Additional File 3: Files S4 and S5). Among onic acid) nor the minor (valeric acid and hexanoic acid, those were metabolites important for methylation and branched short chain fatty acids, 2-methylbutyric of protein lysine and/or carnitine biosynthesis (such acid, isobutyric acid and isovaleric acid) (Additional File as N6-methyllysine, N6, N6, N6-trimethyllysine and 2: Table S5). On the other hand, in the digesta, butyric deoxycarnitine) and microbiota-linked metabolism and valeric acid showed significantly lower values for the (N-methylhydantoin). Supplementation of BC to FOS GOS–BC treatment compared to the control (Table 3). resulted few significant effects (27 and 22 differen- SCFAs in the digesta did not significantly change either tially abundant metabolites respectively in digesta and with the addition of BC to FOS diet or replacement of mucosa samples), generally scattered over the meta- FOS with the GOS in the FOS–BC diet. bolic map, not showing clear effects on any metabolic pathway. On the other hand, replacement of FOS in the Associations between gut microbiota and metabolites FOS–BC diet with GOS, significantly altered a high Correlation analysis number of metabolites in both digesta and plasma (109 The Spearman correlation analysis showed significant and 52 differentially abundant metabolites respectively differences in specific microbe–metabolite correlations in digesta and plasma samples). Unique for the GOS– between the treatments. In the correlation analyses 436 BC treatment were high levels of long chain saturated, digesta metabolites with the human metabolome data- monounsaturated, and polyunsaturated fatty acids, base (HMDB) IDs were included. The circos plot and the as well as of branched fatty acids, most pronounced heat map for microbe–metabolite correlations in digesta for digesta (Additional File 3: File S4). Among those samples from comparisons between FOS–BC and FOS, metabolites were n−3 (eicosapentaenoic acid, EPA, and GOS–BC and FOS–BC treatments are presented in 20:5n−3 and docosapentaenoic acid, DPA, 22:5n−3) Fig. 8. Heatmaps show expansion of the results shown in and n−6 fatty acids (linoleic acid, 18:2n−6, eicosa- the circos plots. In the heatmaps, statistically significant dienoic acid, 20:2n−6, arachidonic acid, 20:4n−6, results (p < 0.05) are indicated with asterisks. Correlations adrenic acid, 22:4n−6 and dihomo-gamma-linolenic values (R) and p values for the specific microbe–metabo - acid; 20:3n−6, DPA, 22:5n−6 and tetracosahexaenoic lite correlations are presented in Additional File 4. acid, 24:6n−3). The GOS–BC fed fish also showed Regarding the results of associations analysis of gut increased levels in the digesta of acetylcarnitine, prop- microbiota and metabolite for the comparison between ionylcarnitine, butyrylcarnitine compared to FOS–BC FOS–BC and FOS treatments, circos plot showed that 4 fed fish, as well as compared to the other treatments. different classes of metabolites, carbohydrates, cofactors Levels of several sphingomyelins, ceramides and hexo- and vitamins, amino acids, and lipids were closely corre- sylceramides were also increased distinctively in the lated with genera belonging to Firmicutes, Actinobacte- GOS–BC fed fish compared to the FOS–BC fed fish ria, Proteobacteria and Epsilonbacteoeota phyla (Fig. 8a). and fish from the other treatments. As shown in the heatmap (Fig. 8c), the 15 genera which Table 3 SCFA concentrations in distal intestinal digesta of the fish from four treatments SCFA concentrations in digesta of distal intestine (ng/ml) Control FOS FOS–BC GOS–BC Acetic acid 8.4E+04 ± 3.9E+04 8.3E+04 ± 2.6E+04 6.5E+04 ± 1.8E+04 1.6E+05 ± 7. 5E+04 a ab ab b Butyric acid 83 ± 11 68 ± 7 60 ± 4 54 ± 3 Propionic acid 121 ± 17 101 ± 9 87 ± 7 88 ± 6 a ab ab b Valeric acid 41 ± 4 34 ± 3 32 ± 7 28 ± 1 Hexanoic acid 254 ± 22 225 ± 13 213 ± 1 203 ± 7 2‑Methylbutyric acid 21 ± 4 16 ± 2 16 ± 2 12 ± 1 Isobutyric acid 25 ± 3 20 ± 1.5 21 ± 2 19 ± 1 Isovaleric acid 13 ± 2 14 ± 1 14 ± 1 12 ± 1 Mean value ± SEM are presented for n = 8 samples. Different letters among values indicate statistically significant differences (q ≤ 0.05). Values sharing the same letters are not statistically significant Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 10 of 22 increased in the FOS–BC treatment compared to FOS in FOS–BC group distinguishing it from the FOS group. treatment showed correlation with 12 significantly VIP plot identified 115 separating factors (20 microbial changed metabolites from the same comparison and genera and 95 metabolites) in the GOS–BC group com- found a number of significantly positive correlations pared to the FOS–BC group (Additional File 5: File S2). (between 10 and 11). Genus Pediococcus showed positive Further, Genus Pediococcus was identified as an impor - and significant associations with 10 metabolites includ - tant variable in both FOS–BC and GOS–BC treatments ing lactose, ergosterol, chiro-inositol and ribose (Addi- from the control and the FOS treatment. Among the tional File 4: File S1). metabolites, lactose, ergosterol and deoxycarnitine found The comparison between the GOS–BC and FOS–BC to be separating factors for FOS and FOS–BC groups, as treatments showed the highest number of associations well as GOS–BC and FOS–BC groups. between microbiota and metabolites, and most of them were significant. Circos plot showed that seven differ - Discussion ent classes of metabolites including nucleotides, carbo- Eec ff ts of supplementation with P. acidilactici to the FOS hydrates, peptides, cofactors and vitamins, xenobiotics, diet amino acids, and lipids were closely correlated with gen- Improved growth was observed for fish fed the FOS and era mainly belonging to Firmicutes, Actinobacteria and P. acidilactici diet when compared to the commercial Proteobacteria phyla (Fig. 8b). As shown in the heatmap control diet. However, it is not possible for us to evaluate (Fig. 8d), all the 24 decreased genera in GOS–BC treat- whether the synbiotic treatment was the causative fac- ment compared to FOS–BC treatment displayed positive tor for the observed improvements in growth, since the correlation with several metabolites (between 54 and 56 experimental diets also contained elevated levels of vita- metabolites). On the other hand, those genera showed min C and E, beta glucan and nucleotides, and had a par- negative correlations with n−3 and n−6 polyunsaturated tial substitution of standard fish meal with krill meal. A fatty acids (Additional File 4: File S2). previous study reported no significant change in growth when Atlantic salmon were fed a diet supplemented with Supervised multivariate analysis the same synbiotic combination [7]. As such, it is possible Supervised multivariate analysis on the combined data that the improved growth observed in the current study matrix of microbiota (at genus level) and metabolites in was caused by other dietary supplements besides the syn- the digesta with the OPLS-DA method pointed out some biotics, or by a combined effect. separation between FOS–BC and FOS treatments as The observation in the present study showing that indicated by the first component (Additional File 1: Fig. alteration in diet composition, in this case supplemen- S4a). On the other hand, it showed a clear separation tation with P. acidilactici to the FOS containing diet, between GOS–BC and FOS–BC treatments (Additional modified the digesta-associated microbiota in the distal File 1: Fig. S4b). intestine of post-smolt Atlantic salmon is in line with Variable importance plot (not shown) based on the other recent observations in salmon [25]. The same is OPLS-DA model was used to identify differential the case for the results regarding the mucosa-associated microbes and metabolites contributing to the separation microbiota, which showed resistance towards dietary of one group compared to the other (Variable Impor- changes, again confirming the results from Li et al. [25] tance on Projection, VIP values > 1 and correlation coef- and supported by the findings from Abid et al. [7]. More - ficients p < 0.05). The list of microbiota and metabolites over, our findings that, irrespective of diet, genera Lac - fulfilling the above statistical criteria in each comparison tobacillus and Leuconostoc dominated in the digesta and between dietary treatments are presented in the Addi- Brevinema, Aliivibrio and Lactobacillus dominated in tional File 5. The FOS–BC treatment showed 41 sepa - mucosa are also strengthening previous findings which rating factors (19 microbial genera and 22 metabolites), indicate that these bacterial groups are among the core from the FOS treatment (Additional File 5: File S1). The microbiota in digesta and mucosa in post-smolt Atlantic genus Pediococcus was identified as an important variable salmon [25–27]. Alpha diversity, or species richness of (See figure on next page.) Fig. 8 Circos plots (a, b) and heatmaps (c, d) showing associations analysis. Associations analysis performed between differentially abundant microbiota and metabolites in FOS–BC group compared to FOS group (a, c) and GOS–BC group compared to FOS–BC group (b, d) based on Spearman Correlation Analysis. Heatmaps show expansion of the results shown in the circos plots. Spearman’s correlation, R, ranges between − 1 to 1. p < 0.05 indicates a statistically significant correlation. In circos plots, red and green lines specify positive and negative correlations, respectively. In heatmaps, red color and blue color indicate positive and negative correlations respectively. The darker color indicates the larger statistical significance. Symbol * and ** indicate p value for correlation coefficients smaller than 0.05 or 0.01, respectively. Correlations between differential microbiota and metabolites among the treatments including R and p values are presented in Additional File 4 Actinobacteria Actinobacteria Epsilonbacteraeota Firmicutes Firmicutes Proteobacteria Proteobacteria Savagea Ruminococcaceae UCG-012 Staphylococcus Vagococcus Tessaracoccus Gallicola Pisciglobus Brachybacterium Geobacillus Enterococcus Dietzia Aneurinibacillus Arthrobacter Rhodococcus Acetobacter Methylococcus Arcobacter Thermomonas Exiguobacterium Gordonia Mobilicoccus Afipia ML602J-51 ML602J-51 Jeotgalibaca Aerococcus Proteus Saccharopolyspora Kocuria Solobacterium Micrococcus Pediococcus Pisciglobus Stenotrophomonas Actinomyces Ochrobactrum Streptolococcus Gordonia Kurthia Peptococcus Staphylococcus Peptostreptococcus Tissierella Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 11 of 22 Phylum Phylum Fig. 8 (See legend on previous page.) Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 12 of 22 the digesta-associated microbiota in the fish fed FOS–BC [30] and juvenile rockfish, Sebastes schlegeli [6], whereas diet did not differ significantly from those fed FOS diet a more recent study reported no significant changes in indicating that supplementation of P. acidilactici to the growth parameters in rainbow trout upon application of FOS diet did not alter the number of ASVs in digesta. the same synbiotic combination in rainbow trout diets Similar to the present observation, a previous study [7] [19]. The results of the present study are in line with the has reported that dietary application of FOS and P. acidi- results of the latter work. lactici did not induce significant changes in alpha diver - sity in the digesta-associated microbiota in the distal Effects on microbiota intestine of Atlantic salmon. On the other hand, the fact Replacing FOS with GOS induced a reduction in alpha that fish fed FOS–BC diet showed a significantly differ - diversity i.e. a decrease in the number of ASVs present, ent beta diversity in the digesta from that of the fish fed as well as a change in beta diversity indicating a shift of FOS diet, indicates an ability of P. acidilactici when in abundance of some of the bacteria. Fish in the GOS– combination with FOS to modulate the bacterial compo- BC treatment showed increased relative abundance sition by altering abundance of the various bacteria. As of P. acidilactici compared to the FOS–BC treatment. expected, P. acidilactici showed relatively higher abun- This increase in both the digesta and mucosa indicates dance in both the digesta and mucosa samples of fish fed enhanced establishment of P. acidilactici when in com- FOS–BC diet compared to those fed FOS diet, suggesting bination with GOS relative to FOS. Increased abundance strengthened establishment in the gut. Previous studies of P. acidilactici was also reported in Rainbow trout have demonstrated the ability of P. acidilactici to popu- treated with GOS in combination with the same pro- late the distal intestine of Atlantic salmon when used biotic species [19]. Pediococcus acidilactici strains are as dietary supplementation [10] or in combination with known to produce bacteriocins, pediocins, which may FOS [7]. In the present study, P. acidilactici was found exert antagonistic effects towards a variety of bacteria to be the key factor separating gut microbiota of fish fed including both gram negative and positive species [31, FOS–BC diet from that of the fish fed FOS diet indicat - 32]. Therefore, decreased alpha diversity and the reduced ing its importance for modulating the gut microbiota abundance of several genera observed when replacing profile. The P. acidilactici was also the key factor for dis - FOS with GOS in the diet could potentially be a result tinguishing the digesta-associated microbiota profiles of of increased antagonistic effects exerted by P. acidilac - FOS–BC fed fish from those in the fish fed control diet. tici when combined with GOS. However, this should be Since gut microbiota plays an important role in shap- further investigated with measurements of pediocins in ing the fecal metabolome, we expected that the com- the digesta, mucosa and blood samples as reduced rela- bined evaluation of microbiota and metabolomics data tive abundance of other genera could also simply be due could give us a better understanding of the possible func- to the increased relative abundance of P. acidilactici. tional implications of the diets used in the present study. We analyzed the global metabolite signature and SCFA Impact on the metabolome of digesta and blood plasma levels in both digesta and blood plasma, expecting that Replacing FOS with GOS increased levels of short, microbiota and metabolic associations may give local medium, and long chain acyl-carnitines in both digesta as well as systemic effects [28, 29]. However, only a few and plasma. This suggests that GOS could directly influ - metabolites showed significant difference between the ence or act as a substrate for the gut microbiota to supply fish fed FOS–BC and FOS diets, demonstrating that the the intestinal mucosa and the body with compounds hav- supplementation of P. acidilactici to the FOS diet had a ing important functions in lipid transport and metabo- minor impact on the metabolome of both the gut content lism. Carnitine and its acyl esters (acyl-carnitines) are and the systemic circulation. As a consequence, asso- essential for transport of fatty acids across the outer and ciations between gut microbiota and metabolites were inner mitochondrial membranes, for the mitochondrial also few. The general lack of strong host responses after beta-oxidation of long-chain fatty acids, as well as for P. acidilactici supplementation to the FOS diet was also maintenance of the ratio of acetyl-CoA/CoA [33, 34]. On observed on the transcriptional level. the other hand, gut bacteria can utilize carnitine for pro- tection against osmotic stress [35]. Eec ff ts of replacement of GOS for FOS in the FOS–BC diet The increase in several sphingolipids, including sphin - No previous published studies have compared GOS– gomyelin and interrelated products such as ceramide, BC supplementation with FOS–BC supplementation to and hexosylceramides in the fish fed GOS–BC diet sug - fish feeds. However, supplementation of GOS–BC to a gests that GOS may affect and possibly improve various control diet was previously reported to increase growth barrier functions. Sphingolipids, mainly ceramide, act as performance and lower FCR in rainbow trout fingerlings signaling molecules and are involved in diverse processes Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 13 of 22 including epithelial integrity, cell growth and death, the increase in expression of transcripts (duox, duox2, apoptosis, immunity, and inflammation [36, 37]. As they noxo1a and noxo1b) important for reactive oxygen spe- are important in orchestration of immune responses cies generation and innate host defense pathways on (cytokine release, inflammatory responses and initiation mucosal surfaces, cellular signaling, regulation of gene of apoptosis of the infected cell) and eliminating invad- expression and cell differentiation [47, 48]. The GOS–BC ing pathogens [37], many pathogens have developed treatment also displayed increased expression of the key strategies to exploit host cell sphingolipid pathways to antioxidant enzyme, gpx1b, involved in protection of the change the sphingolipid balance to facilitate their colo- fish from oxidative stress. nization [38]. Therefore, it is possible that the increase Upregulation of genes involved in immune and in sphingolipid levels in GOS–BC fed fish might trigger other defense mechanisms does not necessarily mean an immune response. The transcriptome results seem to increased resistance towards infection diseases or other indicate such an effect as explained below. stressors—it could also representant an adaptation to the Most of the SCFAs levels were quite similar in the fish diet without important implications for disease resist- of the four treatments. The exceptions were butyric acid ance. Before conclusions regarding effects on robustness and valeric acid which showed a reduction in fish fed of the fish can be made, follow-up studies involving infec - GOS–BC diet compared to those fed control. SCFAs are tious challenge or other stress challenge studies should among the most important microbial metabolites in the be conducted. The effects on immune and other defense gut and are reported to exert multiple beneficial effects genes can also possibly be due to an increase in produc- on vertebrates by involvement in energy homeostasis tion of pediocin with antimicrobial properties by the and healthy immune responses [39, 40]. However, only a highly abundant P. acidilactici when in combination with few studies investigating pre- pro- or synbiotic applica- GOS. On the other hand, activation of defense mecha- tions in fish have reported effects on SCFA production, nisms may also be a sign of inflammatory responses. and the observations are quite different from those in the However, the histological appearance of the distal intes- present study. For example dietary application of Ente- tine did not indicate altered state of inflammation which rococcus faecalis in Javanese carp, Puntius gonionotus was evaluated as mild to moderate for all treatments. increased the intestinal propionic and butyric acid, but The mechanism underlying the alteration in the tran - not acetic acid [41]; and Alcaligenes sp. increased intes- scriptome may be the combined effects of (a) the direct tinal acetic acid, but not butyric acid levels in Malaysian influence of GOS, and (b) the indirect influence caused Mahseer, Tor tambroides [42]. Mammalian studies have by the action of microbiota on the GOS and (c) effects of indicated that formation of SCFAs by intestinal bacteria altered metabolite production in the microbiota linked to is regulated through many different host, environmen - the alteration in beta diversity. Support for the suggestion tal, dietary and microbiological factors with substrate of beneficial effect of GOS on disease resistance is found availability, microbial species composition and intestinal in studies with rainbow trout in which a combination of transit time playing a larger role [43]. Therefore, relatively GOS and P. acidilactici increased antioxidant defense similar SCFA concentrations observed among the fish in biomarkers, innate immune responses, and resistance to the three treatments and control possibly indicate that streptococcosis [17, 18]. SCFA regulation is quite stable in the Atlantic salmon even if the dietary and microbial compositions differed Correlations between impacts on microbiota, metabolome among the treatments. However, this needs to be further and transcriptome investigated. Replacing FOS with GOS in the FOS–BC diet showed significant impacts on gut microbiota and metabolite Impact on the transcriptome associations. Spearman correlation analysis revealed that The observed transcriptomic changes upon the FOS to metabolites including nucleotides, carbohydrates, pep- GOS exchange, i.e. upregulation of genes coding for a tides, cofactors and vitamins, xenobiotics, amino acids, number of cytokines and/or their receptors (Il17a, il17a/ and lipids were closely correlated with genera mainly f1, i17ra, tnfrsf1b, tnfrsf1, tnfrsf9a, tnfsf18 and il-1rl) belonging to Firmicutes, Actinobacteria and Proteobac- indicate alterations in communication between innate teria phyla. Previous studies in fish and mammals have and adaptive immune systems [44, 45]. The increase reported the involvement of gut microbiota in lipid in expression of the toll-like receptor 18 gene (tlr18), metabolism and energy homeostasis [49, 50] and de novo important for bacterial pathogen recognition [46], and synthesis of essential amino acids and vitamins [51, 52]. of the antibacterial peptide gene (hepc1) may indicate This suggests that supplementation of GOS and P. acidi - effects of the exchange of prebiotic on immune func - lactici in the diet could have modulated gut microbiota tions important for disease resistance. The same regards associated with some of those functions in the post-smolt Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 14 of 22 Atlantic salmon in the present study as well. Further, with krill meal. The experimental diets were further sup - increased transcripts and metabolite levels related to plemented with either; prebiotic fructo-oligosaccharide immunomodulatory effects could also potentially link to (FOS, 0.1%), FOS (0.1%) and Bactocell (0.03%) (FOS–BC); the increased abundance of P. acidilactici when in combi- or galacto-oligosaccharide (1.0%) and Bactocell (0.03%) nation with GOS. (GOS–BC) (Table 4). Bactocell (Lallemand Inc., Cardiff, UK) is authorized by the European Union for the use in fish and shrimp [53] and has already been used in salmon Conclusions fry and freshwater stage diets. All feeds were produced This study reports effects on growth performance, gut at Biomar Feed technology Center in Brande, Denmark. health, microbiota, transcriptome, metabolome, and Four randomly distributed pens were allocated for each their associations in post-smolt Atlantic salmon fed diets dietary group. Fish were fed above mentioned four feeds: containing the prebiotic FOS, a combination of FOS and acclimatization diets (3, 5 mm pellets) during the first the probiotic P. acidilactici, or a combination of GOS 5 weeks following seawater transfer, and then the trial and P. acidilactici. No significant effects of these dietary diets (5 mm pellets) for 10 additional weeks. Bulk weights alterations were detected on growth or histomorpho- for each pen were registered at the end of the acclima- logical appearance of the gut. Supplementation with P. tization period and the 10-week feeding trial period to acidilactici to the FOS containing diet altered digesta determine start and end weight of the experimental fish. associated microbiota to some degree, whereas the During the experimental period, average seawater tem- mucosa-associated microbiota seemed relatively resistant perature of 12.4 ± 1.8 °C, salinity of 31.9 ± 0.7 ppt and to such dietary modulation. This probiotic also induced oxygen of 10.0 ± 1.1 mg/l were reported. moderate effects in some of the assessed components At the end of the feeding trial, four fish were ran - of the metabolome and transcriptome. Replacing FOS domly taken from each net pen, anesthetized with tric- with GOS in FOS–BC diet induced several, clear effects aine methanesulfonate (MS222 ; Argent Chemical on many of the observed biomarkers which may indicate Laboratories, Redmond, WA, USA), weighed individu- that GOS induces important effects on the microbiota, ally and euthanized by a sharp blow to the head. Blood metabolome in the digesta as well as the endogenous metabolism, as well as on the mucosal metabolism and function. However, those alterations did not significantly impact the growth performance of GOS–BC group. Fur- Table 4 Composition of experimental diets for post‑smolt ther infection challenge and stress studies are needed to Atlantic salmon ascertain the efficacy of dietary application of GOS and P. Diet composition (g/100 g) Trial feeds (5 mm pellet size) acidilactici along with functional ingredient mixes as an Control FOS FOS–BC GOS–BC immune stimulant strategy against disease outbreaks and stressful events. Fish meal 15.0 15.0 15.0 15.0 Soya SPC 11.0 11.0 11.0 11.0 Materials and methods Wheat Gluten 7.2 8.0 8.0 8.0 Experimental design, study parameters and analytical Maize gluten 5.0 5.0 5.0 5.0 procedures used to evaluate the effect of functional sea - Pea protein 15.0 15.0 15.0 15.0 water transfer diets for Atlantic salmon are illustrated in Guar meal 8.0 7.0 7.0 7.0 the Fig. 1 and explained in the subsequent sections. Wheat 11.0 10.8 10.9 10.0 Fish oil 13.2 11.5 11.5 11.5 Feeding trial Rapeseed oil 10.4 11.1 11.1 11.1 A sea water feeding trial was conducted with post-smolt Vit + min + AA 4.3 4.9 4.9 4.9 Atlantic salmon at LetSea research facility in Dønna, Yttrium 0.1 0.1 0.1 0.1 Norway from 29/05/2018 to 16/09/2018, following the FOS – 0.1 0.1 – Norwegian laws regulating the experimentation with live GOS – – – 1.0 animals. Bactocell 0.03 0.03 Atlantic salmon with average weight 172 ± SEM 0.89 g Water change − 0.1 0.5 0.5 0.5 were randomly assigned to 16 net pens (5 × 5 × 5 m) with Analyzed moisture (%) 5.8 5.4 5.7 6 300 fish each. Four feeds were prepared by Biomar AS, a Energy (bomb calorimetry, MJ/kg) 24.2 24.2 23.8 24.1 control diet based on standard grower feed recipes and Crude FAT (%) 28.5 27.9 27.6 28.1 three experimental diets. The experimental diets con - Crude protein (%) 43.2 43.5 43.9 43 tained elevated vitamin C and E, beta glucan and nucleo- Beta glucan, nucleotides and krill were added only to the experimental diets in tides, and had a partial substitution of standard fish meal equal amounts Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 15 of 22 samples were drawn from the caudal vein using heparin- The degree of change was graded using a scoring sys - ized syringes and placed on ice before plasma collection. tem with a scale of 0–4 where 0 represented normal; 1, Plasma was collected after centrifugation at 2000g for mild; 2, moderate; 3, marked, and 4, severe changes. The 10 min (4 °C) and snap frozen in liquid N2. After clean- histological evaluation was conducted randomly and ing the exterior of each fish with 70% ethanol, the distal blind, and assignment of individual samples to the treat- intestine was aseptically removed, opened longitudinally ments was obtained after the evaluation was completed. and digesta was collected into a 50 ml sterile centrifuge Differences in histological scores for the evaluated tube. The digesta was mixed thoroughly with a spatula morphological characteristics of the intestinal tissue were and aliquots were transferred into 1.5 ml sterile Eppen- analyzed for statistical significance using ordinal logistic dorf tubes and snap frozen in liquid N and stored at regression run in the R statistical package (version 3.6.3; − 80 °C for the analysis of the digesta-associated intesti- 2020) within the RStudio interphase (version 1.3.1093; nal microbiota and metabolomic profiling. The mid-sec - 2020). Differences were examined based on odds ratios of tion of the same distal intestine was excised and rinsed 3 the different treatments having different histology scores times in sterile phosphate-buffered saline. Subsequently, compared to the reference diet. Control was used as the the tissue was transversely divided into 3 pieces, respec- reference. tively, for histological evaluation (fixed in 4% phosphate- buffered formaldehyde solution for 24 h and transferred Microbiota analysis to 70% ethanol for storage), RNA-Sequencing (preserved DNA extraction in RNAlater solution and stored at − 20 °C) and mucosa- For analysis of the distal intestinal microbiota, a total associated intestinal microbiota analysis (snap frozen in of 32 fish samples were used. Two fish were randomly liquid N and stored at − 80 °C). 2 selected from each of the 4 pens allocated for a dieatary The performance of the fish in each dietary group was group to have n = 8 fish per diatary group. The DNA calculated using the thermal growth coefficient and spe - was extracted from respective digesta and mucosa sam- cific growth rate, which are considered as good predic - ples following the protocol of QIAamp Fast DNA Stool tors of salmon growth [54]. Statistical analysis of growth Kit (Qiagen, Crawley, UK) with some modification as parameters among the treatments was performed by suggested by Knudsen et al. [55]. Samples were pre-pro- one-way ANOVA after checking the fulfillment of all cessed with a bead-beating protocol of three times in the the pertinent assumptions, normality of the distribution Fastprep at 6.5 m/s for 30 s with a mix of beads (120 mg and homogeneity of variances. Pairwise comparisons acid-washed glass beads (150–212 μm) and 240 mg Zir- were analyzed using Tukey’s honestly significant different conium oxide beads (1.4 mm). For quality control of (HSD) test, and q ≤ 0.05 was considered as statistically the microbiota profiling protocol, along with the each significant. of the DNA extraction batch, two ‘blanks’ (without any sampling materials) and two ‘positive controls’ i.e. mock (microbial community standard from Zymo-BIOMICS , Histological analysis Zymo Research, California, USA) were included. The The gut tissue sections (total of 64 fish, n = 16 per dietary mock contains 8 bacteria (Pseudomonas aeruginosa, group, n = 4 fish randomly selected from each of the 4 Escherichia coli, Salmonella enterica, Lactobacillus fer- pens allocated for a dietary group) of pyloric caeca and mentum, Enterococcus faecalis, Staphylococcus aureus, distal intestine were evaluated by light microscopy with Listeria monocytogenes, Bacillus subtilis) and 2 yeasts focus on the characteristic morphological changes of soy- (Saccharomyces cerevisiae, Cryptococcus neoformans). bean meal-induced enteritis (SBMIE) in Atlantic salmon distal intestine, that consist of shortening of mucosal fold PCR amplification of V1–V2 region of the 16S rRNA gene length, increase in width and inflammatory cell infiltra - PCR amplification was carried out using 27F (5′ AGA tion of the submucosa and lamina propria, and reduction GTT TGATCMTGG CTC AG 3′), and 338R-I (5′ GCW in enterocyte supranuclear vacuolization. Additionally, GCC TCC CGT AGG AGT 3′) and 338R-II (5′ GCW GCC for the pyloric caeca, changes in the vacuolization of the ACC CGT AGG TGT 3′) to have about 300 bp ampli- intestinal epithelial cells were evaluated. Normally, little cons [26]. PCRs were carried out in 25 μl reactions to no vacuolization is present in the intestinal epithelial ® with 12.5 μl of Phusion HighFidelity PCR Master Mix cells of the pyloric caeca and mid intestine. Increased (Thermo Scientific, CA, USA); 1 μM of forward and vacuolization (or hyper-vacuolization) is observed in fish reverse primers, and 1 μl template DNA. Undiluted and affected by the so-called lipid malabsorption syndrome 1:2 diluted templates were used, respectively, from the (LMS) that manifests in its advanced form as ‘floating digesta and mucosa. The PCR conditions were as follows: feces’ (steatorrhea). initial denaturation at 98 °C for 7 min followed by initial Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 16 of 22 10 cycles with denaturation at 98 °C for 30 s, annealing trimmed and quality filtered using the DADA2 algorithm temperature decreasing from 63 to 53 °C for 30 s at each [62] in QIIME2. Primer sequences were trimmed off (for - temperature and extension at 72 for 30 s; followed by 25 ward reads, first 20bps; reverse reads, first 18bps) and the further cycles with denaturation at 98 °C for 30 s, anneal- reads were truncated at the position where the median ing at 53 °C for 30 s, and extension at 72 °C for 30 s; fol- Phred quality crashed (forward reads, at position 290 bp; lowed by a final extension at 72 °C for 10 min. Negative reverse reads, at position 238 bp) and low-quality reads PCR controls were included by replacing the template were filtered out. Chimeric sequences were removed DNA with molecular grade water. PCR was performed after merging the reads. The taxonomy was assigned to in duplicate, pooled, and examined by 1.5% agarose gel resulting amplicon sequence variants (ASVs) tables by a electrophoresis. Scikitlearn Naive Bayes machine-learning classifier [63], which was trained on the SILVA 132 99% ASVs [64] that Library preparation and sequencing were trimmed to exclusively include the regions of 16S Library preparation of the products from amplicon PCR rRNA gene amplified by the primers used in the cur - was performed using the Quick-16S NGS Library Prep rent study. Filtering of ASVs table was performed using Kit (Zymo Research) following the instructions from the q2-feature-table plugin in Qiime2. ASVs assigned as producer. Briefly, PCR products were first enzymatically chloroplast and mitochondria were removed from ASVs cleaned up followed by a PCR to add barcodes. Subse- table. The ASVs table was then filtered to remove ASVs quently, the libraries were quantified by qPCR, pooled, that were without a phylum-level taxonomic assignment and purified. A representative number of individual or appeared in only one biological sample. Low abun- libraries were evaluated for DNA quality in Agilent Bio- dance ASVs with total abundance of less than 2 across analyzer 2100 system (Agilent Technologies, California, all the samples were also filtered out. Contaminant USA). The final pooled library was then denatured and sequences were detected using control samples (nega- diluted to 8 pM and sequenced on Illumina MiSeq plat- tive PCR reactions, DNA extraction blanks and mocks) form with Miseq Reagent Kit v3 (600-cycle) (Illumina) to and bacterial DNA quantification data obtained from generate paired-end read. 20% of 8 pM PhiX control was qPCR mentioned in the previous section, as suggested added as an internal control. by Davies et al. [65]. In general, contaminants are fre- quently found in negative controls and blanks and show Bacterial DNA quantification by qPCR a negative correlation with the bacterial DNA concentra- As an extra measure to identify contaminating tion. Moreover, contaminants also can be foreign ASVs sequences, qPCR was performed to quantity 16S rRNA in mocks those are not belonging to the original included gene in the diluted DNA templates (samples, blanks, and bacteria. In total 17 and 11 ASVs were removed from mocks) used for the amplicon PCR. The qPCR assays mucosa and digesta samples respectively based on their were performed using a universal primer set (forward, presence in mocks, extraction blanks and negative PCR 5′-CCA TGA AGT CGG AAT CGC TAG-3′; reverse, controls, and their negative correlation with bacterial 5′-GCT TGA CGG GCG GT G T-3′) as described previ- DNA concentration. The ASVs removed from mucosa ously [56, 57]. The qPCR was performed using the Light - samples belonged to the genera Rhodoluna (1 ASV), Cycler 96 (Roche Applied Science, Basel, Switzerland) Cutibacterium (1 ASV), Flavobacterium (6 ASVs), Afipia in a 10 µl reaction volume; 2 µl of PCR-grade water, 1 µl (1 ASV), Curvibacter (2 ASVs), Limnohabitans (1 ASV), diluted DNA template, 5 µl LightCycler 480 SYBR Green Polynucleobacter (1 ASV), Ralstonia (2 ASVs), Undibac- I Master Mix (Roche Applied Science) and 1 µl (3 µM) of terium (1 ASV) and Pseudomonas (1 ASV). On the other each primer. The qPCR program used as follows; an ini - hand, the removed contaminants from digesta samples tial enzyme activation step at 95 °C for 2 min, 45 three- belonged to the genera Flavobacterium (6 ASVs), Curvi- step cycles of 95 °C for 10 s, 60 °C for 30 s and 72 °C for bacter (2 ASVs), Rhodoluna (1 ASV), Polynucleobacter (1 15 s, and a melting curve analysis at the end. Quantifi - ASV) and Ralstonia (1 ASV). After filtering, a total num - cation cycle (Cq) values were determined using the sec- ber of 1 075 and 385 ASVs were obtained for digesta and ond derivative method [58] and bacterial DNA standards mucosa samples, respectively. The ASVs filtered from the were used as inter-plate calibrators and the inter-plate raw ASVs table were also removed from the representa- calibration factor was calculated as described previously tive sequences. The final ASVs tables with taxonomy are [59]. presented in Additional File 6. Diversity analysis was performed using q2-diversity Bioinformatics analysis of microbiota sequencing data plugin in Qiime2. To compute alpha and beta diver- This was performed using QIIME2 version 2 [60, 61]. sity indices, the ASVs tables were rarified at 28,295 and The demultiplexed paired-ended reads were denoised, 15,655 reads for digesta and mucosa samples respectively Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 17 of 22 in order to have an even number of reads across all the negative binomial (Gamma-Poisson) distribution. The samples. The rarefaction curves based on observed ASVs analysis is executed through 3 main steps; estimation of for the digesta and mucosa samples from 32 fish and from size factors, estimation of dispersion, and negative bino- each feed group are presented in Additional File 1: Figs. mial generalized linear model fitting and Wald statistics S5 and S6 for digesta and mucosa, respectively. Alpha [72]. DESeq2 uses un-normalized count data as input, diversity was calculated using observed species and and it internally corrects for library size. DESeq2 per- Shannon`s diversity indices at ASVs level. Beta diversity forms independent filtering by removing genes with low was evaluated using Bray–Curtis at ASVs level followed counts which are not likely to produce significant differ - by PERMANOVA analysis along with pairwise compari- ences due to high dispersion. It uses the mean of normal- sons. MicrobiomeAnalyst package [66, 67] was used to ized counts irrespective of the biological conditions for analyze abundant taxa among treatments, Random For- independent filtering [72]. By default, DESeq2 replaces est analysis, NMDS analysis and graphical presentations outliers if the Cook’s distance is large for a sample. Dif- of data using ASVs tables. ferential expression was calculated for pairwise compari- sons using un-transformed data. The differences were Global transcriptomic profiling considered statistically significant when the adjusted p RNA sequencing value (q) with the Benjamini–Hochberg procedure ≤ 0.1. Total RNA was extracted from distal intestinal digesta For the visualization of DEGs in heatmaps, log trans- of 28 fish (n = 7 per dietary group) from the 32 fish used formed count data was used. for microbiota analysis using Invitrogen PureLink RNA Mini Kit with column based purification (Thermo Fisher Functional annotation and gene ontology analysis of DEGs Scientific, Waltham, USA), following the manufacturer’s Functional annotation of the DEGs was performed using protocol. Tissues were homogenized twice at 5000×g for g:Profiler online tool [73, 74] and manually inspect- 15 s with zirconium oxide beads (1.4 mm) using Fast- ing the Ensembl (http:// www. ensem bl. org) and NCBI Prep-24 (MP Biomedicals, Thermo Fisher Scientific, (https:// www. ncbi. nlm. nih. gov/) data bases. Gene ontol- Waltham, USA). RNA integrity was checked using an ogy enrichment analysis (GO) was carried out also with Agilent 2200 TapeStation (Agilent Technologies, Santa g:Profiler online tool. For the calculation of statisti - Clara, USA), and RNA quantity and RNA purity were cally significant enrichment, all the known genes of the measured using Epoch Microplate Spectrophotometer Atlantic salmon in the Ensembl database (Ensembl 100, (BioTeK Instruments, Winooski, USA). Ensemble genome 47) were considered and the threshold Library preparation and RNA sequencing was per- to determine GO terms was set as Benjamini–Hochberg formed by Norwegian National Sequencing Center FDR (False Discovery Rate) value of 0.1. Enriched GO (Oslo, Norway). Libraries were prepared using TruSeq terms were then summarized by removing redundant Stranded mRNA Library Prep kit with TruSeq RNA GO terms and visualized in semantic similarity-based unique dual indexes in accordance with the manufactur- scatterplots using REVIGO online tool [75]. er’s protocol (Illumina, San Diego, USA). Sequencing was performed on the Illumina SP Novaseq flow cell to yield Short chain fatty acids and metabolites analysis 100 bp single end reads. Targeted short chain fatty acids analysis and global untar- geted metabolite profiling were performed by Metabo - Bioinformatics analysis of RNA‑seq data lon, Inc. (Morrisville, USA). Plasma and digesta collected After demultiplexing, raw sequencing data was pro- from the same 32 fish (n = 8 per dietary group) used for cessed for quality and adapter trimming using Cuta- microbiota and transcriptomics analysis. dapt [68] with − q 25, 20, quality-base = 33, trim-n -m 20 parameters, followed by a further quality check with SCFA analysis FastQC (https:// www. bioin forma tics. babra ham. ac. uk/ For the SCFA analysis, samples were spiked with sta- proje cts/ fastqc/). Quality trimmed reads were mapped ble labelled internal standards, homogenized, and to the indexed Atlantic salmon genome, ICSASG v2 with subjected to protein precipitation. An aliquot of the refseq genes using HISAT2 package [69] in Norwegian supernatant was derivatized, then diluted and injected e-Infrastructure for Life Sciences (NeLS) galaxy plat- onto liquid chromatography-tandem mass spectrom- form developed by ELIXIR Norway [70]. HTSeq [71] was etry, LC–MS/MS system (Agilent 1290 LC system, Agi- used to compute gene expression values. Differentially lent Technologies Inc, Santa Clara, USA with AB Sciex expressed genes among the treatments were determined QTrap 5500 system, AB Sciex, Framingham, USA). using DESeq2 [72] using the default parameters. DESeq2 The mass spectrometer was operated in negative mode performs differential expression analysis based on the using electrospray ionization (ESI). The peak area of the Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 18 of 22 SCFA and metabolite data analysis individual analyte product ions was measured against Statistical analysis of changes in SCFA concentrations the peak area of the product ions of the correspond- among the treatments were carried out using one-way ing internal standards. Quantification was performed ANOVA followed by Tukey HSD test after checking for using a weighted linear least squares regression analysis the fulfillment of all pertinent assumptions for ANOVA. generated from fortified calibration standards prepared Changes in SCFAs considered statistically significant immediately prior to each run. LC–MS/MS raw data when q ≤ 0.05. For the metabolites data, originally nor- were collected and processed using AB SCIEX software malized data (normalized to correct the variation due Analyst 1.6.2. Analyte concentrations that fell below to instrument inter-day tuning differences) was rescaled and above the limit of quantitation were removed from to set the median equal to 1. Then missing values were the downstream analysis. From all the SCFAs analyzed, imputed with the minimum. Welch’s t-test which allows only 4 out of the 32 samples were below the quantita- for unequal variances was used to analyze changes in tion for one SCFA, isobutyric acid. metabolite concentrations among the treatments and metabolite concentrations considered statistically signifi - cant when p ≤ 0.05. Global metabolite profiling Samples were prepared by automated Microlab STAR (Hamilton company, Reno, USA) system [76]. Metabo- lon inc. used ultraperformance liquid chromatogra- Correlation analysis of microbiota and metabolites phy-tandem mass spectroscopy, UPLC-MS/MS (UPLC Correlation analysis of microbiota and metabolites from Waters ACQUITY, Milford, USA and Q-Exactive was performed using M2IA online tool [78]. As per the mass spectrometer from Thermo Scientific, Waltham, requirement of the tool, only the metabolites with HMDB USA), for the metabolite analysis. After protein pre- IDs (436 and 293 respectively for digesta and plasma), cipitation, the resulting extract was aliquoted, and two and ASVs table with taxonomic annotations and corre- aliquots were analyzed by separate reverse phase (RP)/ sponding reference sequence file generated from QIIME2 UPLC-MS/MS methods with positive mode using ESI; analysis were used. Data was processed by filtering out one aliquot with RP/UPLC-MS/MS with negative mode both the microbiota and metabolic features with missing using ESI; and one aliquot by hydrophilic interaction values found in more than 80% of samples and the rela- chromatography (HILIC)/UPLC-MS/MS with negative tive standard deviation values less than 30%. Minimum mode using ESI. Several controls were analyzed in con- value was selected to impute missing value for both data cert with the experimental samples including a pooled sets. For data normalization, the relative percentage of matrix sample (and/or a pool of well-characterized features calculated based on the total sum scaling was human plasma) served as a technical replicate through- used. For the pair-wise comparisons of the treatments, out the data set; extracted water samples served as pro- Wilcoxon rank-sum test was used and the p < 0.05 was cess blanks; and a cocktail of QC standards (carefully considered as statistically significant. selected not to interfere with the spiked endogenous Spearman correlation analysis method was selected compound into all the samples) to monitor instrument to analyze correlations between differentially abundant performance and aid in chromatographic alignment. microbiota (genus level) and metabolite concentrations Instrument variability and overall process variabil- in one dietary group compared to the other. Spear- ity were determined respectively by the standards and man correlation analysis method was recommended by spiked endogenous compounds. the developers of M2IA online tool as it outperforms Raw data was extracted, peak-identified and QC other correlation analysis methods due to its overall processed using hardware and software developed performance regarding specificity, sensitivity, similar - by Metabolon [76, 77]. Metabolites were identified ity, accuracy, and stability with different sparsity [79]. by automated comparison of the ion features in the The coefficient values (R) ranged between − 1 and 1 experimental samples to a reference library of chemical and p < 0.05 was considered statistically significant. The standard entries that included retention time, molecu- results were visualized on circos plots and heatmaps to lar weight (m/z), preferred adducts, and in-source frag- identify bacterial genera that were closely related with ments as well as associated MS spectra and were quality different classes of metabolites. controlled and curated to identify true chemical enti- Supervised multivariate analysis first integrates two ties [76, 77]. Peaks were quantified using area-under- data matrix and then identifies differential variables the-curve. Data normalization step was performed to which significantly contribute to the discrimination correct variation resulting from instrument inter-day between two treatments. We selected the orthogonal tuning differences. partial least squares discriminant analysis (OPLS-DA) Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 19 of 22 method in M2IA to identify the microbiota and metab- specific microbe–metabolite correlations in GOS–BC group compared to olites having a significant role in discriminating one the control group. dietary group from the other. Variables of importance Additional file 5: File S1. Variables of importance identified by V ‑plot to discriminate FOS–BC group from the FOS group. File S2. Variables for group separation were identified and clarified with of importance identified by V ‑plot to discriminate GOS–BC group from variable importance plot. Variables with VIP > 1 and the FOS–BC group. File S3. Variables of importance identified by V ‑plot correlation coefficient (corr.coeffs) p < 0.05 were con- to discriminate FOS group from the control group. File S4. Variables of importance identified by V ‑plot to discriminate FOS–BC group from the sidered statistically significant. control group. File S5. Variables of importance identified by V ‑plot to discriminate GOS–BC group from the control group. Supplementary Information Additional file 6: File S1. ASVs table for digesta samples. File S2. ASVs The online version contains supplementary material available at https:// doi. table for mucosa samples. org/ 10. 1186/ s42523‑ 023‑ 00228‑w. Acknowledgements Additional file 1. Figure S1. The absolute bacterial DNA levels quantified The authors would like to thank the employees at LetSea for conducting the by qPCR. DNA levels in digesta samples (a) and mucosa samples (b) from feeding trial and preparations for sampling. We are also grateful to Ellen Hage each of the treatments. n = 8 fish per group. Error bars represent SEM. No at NBMU (Oslo, Norway) for her skillful organization in sample collection and significant differences (p ≤ 0.05) found among the treatments. Figure S2. technical assistance in the laboratory. We would also like to thank Kirsti E. The alpha diversity indices for digesta and mucosa at ASV level. Observed Præsteng at NMBU (Oslo, Norway) for performing 16S rRNA sequencing. ASVs (a) and Shannon indices (b) for digesta and observed ASVs (c) and Shannon indices (d) for mucosa. p values obtained from Kruskal–Wallis Author contributions analysis among the feed groups are presented above each graph. Each Experiment design: TMK, TF, ÅK and AJT. Providing reagents and materials: TF box plot contains 25% and 75% quartiles of the data set respectively and TMK. Laboratory work and data analyses: AD, AJT and EC. Writing, original at the lower and upper ends of the box. The vertical line inside the box draft: AD. Writing, reviewing, and editing: AD, AJT, EC, TMK, TF and ÅK. All indicates the median, and the ends of the whiskers indicate minimum and authors read and approved the final manuscript. maximum values of the data. Black rectangle indicates mean value of the data and dots display values from individual fish. Figure S3. Top 10 most Funding abundant phyla of digesta (a) and mucosa (b) from distal intestine. The This work was supported by the Norwegian Research Council through a samples are grouped by feed groups: Atlantic salmon fed with a control/ research project (GutBiom project, NFR 281807) and BioMar RD, Trondheim, reference diet and three experimental diets: FOS, FOS–BC, and GOS–BC Norway. diets. The mean relative abundance of phyla per feed group is presented on the right side. Figure S4. Orthogonal partial least squares discriminant Availability of data and materials analysis (OPLS‑DA) score plots. OPLS‑DA score plots of the combined 16S rRNA sequencing and RNA‑seq data are publicly available at the NCBI data matrix of metabolome and microbiota in each of the FOS–BC (a) and Sequence Read Archive (SRA) with the accession numbers SUB8676898 and GOS–BC (b) groups compared to FOS and FOS–BC groups, respectively. SUB8572237 respectively, under the Bioproject PRJNA679207. Each dot indicates an individual sample. Figure S5. The rarefaction curves based on observed ASVs for the digesta samples. Rarefaction curves for the digesta samples from 32 fish (a) and each feed group (b). Each Feed Declarations group contains 8 samples. The ASVs table was rarified at 28 295, which is the minimum number of reads detected in the digesta samples. Figure Ethics approval and consent to participate S6. The rarefaction curves based on observed ASVs for the mucosa All experiments involving Atlantic salmon were conducted in agreement with samples. Rarefaction curves for the mucosa samples from 32 fish (a) and the guidelines provided by the Norwegian Animal Research Authority. from each feed group (b). Each Feed group contains 8 samples. The ASVs table was rarified at 15 655 reads, which is the minimum number of reads Consent for publication detected in the mucosa samples. Not applicable. Additional file 2: Table S1. Significantly changed bacterial genera resulted from pairwise comparisons of treatments. Table S2. Random For‑ Competing interests est confusion matrix for digesta‑associated microbiota. Table S3. Random Torunn Forberg is employed by BioMar RD. The remaining authors declare that Forest confusion matrix for mucosa‑associated microbiota. Table S4. Sum‑ the research was conducted in the absence of any commercial or financial marized enriched biological process GO terms produced using REVIGO relationships that could be construed as a potential competing interests. tool for DEGs in GOS–BC group. Table S5. SCFA concentrations in blood plasma from four treatments. Author details Department of Paraclinical Sciences, Faculty of Veterinary Medicine, Norwe‑ Additional file 3: File S1. List of differentially expressed annotated genes gian University of Life Sciences (NMBU), Ås, Norway. Biomar RD, Trondheim, in FOS–BC group compared to the FOS group. File S2. List of differentially Norway. expressed annotated genes in GOS–BC group compared to the FOS–BC group. File S3. List of differentially expressed annotated genes in GOS–BC Received: 11 August 2022 Accepted: 27 January 2023 group compared to the control group. File S4. Detected metabolites in digesta highlighting differential abundance in pairwise comparisons between treatments. File S5. Detected metabolites in plasma highlight‑ ing differential abundance in pairwise comparisons between treatments. Additional file 4: File S1. The specific microbe–metabolite correla‑ References tions in FOS–BC group compared to FOS group. File S2. The specific 1. Merrifield DL, Dimitroglou A, Foey A, Davies SJ, Baker RTM, Bøgwald J, microbe–metabolite correlations in GOS–BC group compared to FOS–BC et al. The current status and future focus of probiotic and prebiotic appli‑ group. File S3. The specific microbe–metabolite correlations in FOS group cations for salmonids. Aquaculture. 2010;302(1):1–18. https:// doi. org/ 10. compared to the control group. File S4. The specific microbe–metabolite 1016/j. aquac ulture. 2010. 02. 007. correlations in FOS–BC group compared to the control group. File S5. The 2. Ringø E, Song SK. Application of dietary supplements (synbiotics and probiotics in combination with plant products and β‑ glucans) in Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 20 of 22 aquaculture. Aquac Nutr. 2016;22(1):4–24. https:// doi. org/ 10. 1111/ anu. 19. Rasmussen JA, Villumsen KR, Ernst M, Hansen M, Forberg T, 12349. Gopalakrishnan S, et al. A multi‑ omics approach unravels metagenomic 3. Ringø E, Olsen RE, Gifstad TØ, Dalmo RA, Amlund H, Hemre G‑I, et al. and metabolic alterations of a probiotic and synbiotic additive in rainbow Prebiotics in aquaculture: a review. Aquac Nutr. 2010;16(2):117–36. trout (Oncorhynchus mykiss). Microbiome. 2022;10(1):21. https:// doi. org/ https:// doi. org/ 10. 1111/j. 1365‑ 2095. 2009. 00731.x.10. 1186/ s40168‑ 021‑ 01221‑8. 4. Song SK, Beck BR, Kim D, Park J, Kim J, Kim HD, et al. Prebiotics as 20. Modanloo M, Soltanian S, Akhlaghi M, Hoseinifar SH. The effects of single immunostimulants in aquaculture: a review. Fish Shellfish Immunol. or combined administration of galactooligosaccharide and Pediococcus 2014;40(1):40–8. https:// doi. org/ 10. 1016/j. fsi. 2014. 06. 016. acidilactici on cutaneous mucus immune parameters, humoral immune 5. Huynh T‑ G, Shiu Y‑L, Nguyen T ‑P, Truong Q ‑P, Chen J‑ C, Liu C‑H. Current responses and immune related genes expression in common carp (Cypri- applications, selection, and possible mechanisms of actions of synbiotics nus carpio) fingerlings. Fish Shellfish Immunol. 2017;70:391–7. https:// doi. in improving the growth and health status in aquaculture: a review. Fish org/ 10. 1016/j. fsi. 2017. 09. 032. Shellfish Immunol. 2017;64:367–82. https:// doi. org/ 10. 1016/j. fsi. 2017. 03. 21. Soltani M, Badzohreh G, Mirzargar S, Farhangi M, Shekarabi PH, Lymbery 035. A. Growth behavior and fatty acid production of probiotics, Pediococcus 6. Rahimnejad S, Guardiola FA, Leclercq E, Ángeles Esteban M, Castex M, acidilactici and Lactococcus lactis, at different concentrations of fructooli‑ Sotoudeh E, et al. Eec ff ts of dietary supplementation with Pediococcus gosaccharide: studies validating clinical efficacy of selected synbiotics on acidilactici MA18/5M, galactooligosaccharide and their synbiotic on growth performance of Caspian roach (Rutilus frisii kutum) fry. Probiot‑ growth, innate immunity and disease resistance of rockfish (Sebastes ics Antimicrob Proteins. 2019;11(3):765–73. https:// doi. org/ 10. 1007/ schlegeli). Aquaculture. 2018;482:36–44. https:// doi. org/ 10. 1016/j. aquac s12602‑ 018‑ 9462‑3. ulture. 2017. 09. 020. 22. FAO. The state of world fisheries and aquaculture. Rome: FAO; 2020. 7. Abid A, Davies SJ, Waines P, Emery M, Castex M, Gioacchini G, et al. Dietary 23. Thorstad EB, Whoriskey F, Uglem I, Moore A, Rikardsen AH, Finstad B. synbiotic application modulates Atlantic salmon (Salmo salar) intestinal A critical life stage of the Atlantic salmon Salmo salar: behaviour and microbial communities and intestinal immunity. Fish Shellfish Immunol. survival during the smolt and initial post‑smolt migration. J Fish Biol. 2013;35(6):1948–56. https:// doi. org/ 10. 1016/j. fsi. 2013. 09. 039. 2012;81(2):500–42. https:// doi. org/ 10. 1111/j. 1095‑ 8649. 2012. 03370.x. 8. Merrifield DL, Bradley G, Harper GM, Baker RTM, Munn CB, Davies SJ. 24. Wang J, Jaramillo‑ Torres A, Li Y, Kortner TM, Gajardo K, Brevik ØJ, et al. Assessment of the effects of vegetative and lyophilized Pediococcus Microbiota in intestinal digesta of Atlantic salmon (Salmo salar), observed acidilactici on growth, feed utilization, intestinal colonization and health from late freshwater stage until one year in seawater, and effects of parameters of rainbow trout (Oncorhynchus mykiss Walbaum). Aquac functional ingredients: a case study from a commercial sized research site Nutr. 2011;17(1):73–9. https:// doi. org/ 10. 1111/j. 1365‑ 2095. 2009. 00712.x. in the Arctic region. Anim Microbiome. 2021;3(1):14. https:// doi. org/ 10. 9. Ferguson RMW, Merrifield DL, Harper GM, Rawling MD, Mustafa S, Pic‑1186/ s42523‑ 021‑ 00075‑7. chietti S, et al. The effect of Pediococcus acidilactici on the gut microbiota 25. Li Y, Bruni L, Jaramillo‑ Torres A, Gajardo K, Kortner TM, Krogdahl Å. and immune status of on‑ growing red tilapia (Oreochromis niloticus). J Differential response of digesta‑ and mucosa‑associated intestinal Appl Microbiol. 2010;109(3):851–62. https:// doi. org/ 10. 1111/j. 1365‑ 2672. microbiota to dietary insect meal during the seawater phase of Atlantic 2010. 04713.x. salmon. Anim Microbiome. 2021;3(1):8. https:// doi. org/ 10. 1186/ 10. Jaramillo‑ Torres A, Rawling MD, Rodiles A, Mikalsen HE, Johansen L‑H, Tin‑s42523‑ 020‑ 00071‑3. sley J, et al. Influence of dietary supplementation of probiotic Pediococcus 26. Gajardo K, Jaramillo‑ Torres A, Kortner TM, Merrifield DL, Tinsley J, Bakke acidilactici MA18/5M during the transition from freshwater to seawater AM, et al. Alternative protein sources in the diet modulate microbiota and on intestinal health and microbiota of Atlantic slmon (Salmo salar L.). functionality in the distal intestine of Atlantic salmon (Salmo salar). Appl Front Microbiol. 2019. https:// doi. org/ 10. 3389/ fmicb. 2019. 02243. Environ Microbiol. 2017. https:// doi. org/ 10. 1128/ aem. 02615‑ 16. 11. Kiron V, Kulkarni A, Dahle D, Vasanth G, Lokesh J, Elvebo O. Recognition of 27. Gajardo K, Rodiles A, Kortner TM, Krogdahl Å, Bakke AM, Merrifield DL, purified beta 1,3/1,6 glucan and molecular signalling in the intestine of et al. A high‑resolution map of the gut microbiota in Atlantic salmon Atlantic salmon. Dev Comp Immunol. 2016;56:57–66. https:// doi. org/ 10. (Salmo salar): a basis for comparative gut microbial research. Sci Rep. 1016/j. dci. 2015. 11. 007. 2016;6:30893. https:// doi. org/ 10. 1038/ srep3 0893. 12. Wang J, Kortner TM, Chikwati EM, Li Y, Jaramillo‑ Torres A, Jakobsen JV, 28. Vernocchi P, Del Chierico F, Putignani L. Gut microbiota profiling: metabo ‑ et al. Gut immune functions and health in Atlantic salmon (Salmo salar) lomics based approach to unravel compounds affecting human health. from late freshwater stage until one year in seawater and effects of func‑ Front Microbiol. 2016. https:// doi. org/ 10. 3389/ fmicb. 2016. 01144. tional ingredients: a case study from a commercial sized research site in 29. Holmes E, Li JV, Athanasiou T, Ashrafian H, Nicholson JK. Understanding the Arctic region. Fish Shellfish Immunol. 2020;106:1106–19. https:// doi. the role of gut microbiome–host metabolic signal disruption in health org/ 10. 1016/j. fsi. 2020. 09. 019. and disease. Trends Microbiol. 2011;19(7):349–59. https:// doi. org/ 10. 13. Gibson GR, Roberfroid MB. Dietary modulation of the human 1016/j. tim. 2011. 05. 006. colonic microbiota: introducing the concept of prebiotics. J Nutr. 30. Hoseinifar SH, Mirvaghefi A, Amoozegar MA, Merrifield DL, Ringø E. 1995;125(6):1401–12. https:// doi. org/ 10. 1093/ jn/ 125.6. 1401. In vitro selection of a synbiotic and in vivo evaluation on intestinal 14. Markowiak P, Śliżewska K. Eec ff ts of probiotics, prebiotics, and synbiotics microbiota, performance and physiological response of rainbow trout on human health. Nutrients. 2017;9(9):1021. (Oncorhynchus mykiss) fingerlings. Aquac Nutr. 2017;23(1):111–8. https:// 15. Aftabgard M, Salarzadeh A, Mohseni M. The effects of a synbiotic doi. org/ 10. 1111/ anu. 12373. mixture of galacto‑ oligosaccharides and Bacillus strains in Caspian 31. Papagianni M, Anastasiadou S. Pediocins: The bacteriocins of Pediococci. salmon, Salmo trutta caspius Fingerlings. Probiotics Antimicrob Proteins. Sources, production, properties and applications. Microb Cell Fact. 2019;11(4):1300–8. https:// doi. org/ 10. 1007/ s12602‑ 018‑ 9498‑4. 2009;8:3. https:// doi. org/ 10. 1186/ 1475‑ 2859‑8‑3. 16. Hoseinifar SH, Yousefi S, Van Doan H, Ashouri G, Gioacchini G, Maradonna 32. Porto MC, Kuniyoshi TM, Azevedo PO, Vitolo M, Oliveira RP. Pediococcus F, et al. Oxidative stress and antioxidant defense in fish: the implications spp.: an important genus of lactic acid bacteria and pediocin producers. of probiotic, prebiotic, and synbiotics. Rev Fish Sci Aquac. 2020;29:1–20. Biotechnol Adv. 2017;35(3):361–74. https:// doi. org/ 10. 1016/j. biote chadv. https:// doi. org/ 10. 1080/ 23308 249. 2020. 17956 16.2017. 03. 004. 17. Hoseinifar SH, Mirvaghefi A, Amoozegar MA, Sharifian M, Esteban MÁ. 33. Bremer J. Carnitine–metabolism and functions. Physiol Rev. Modulation of innate immune response, mucosal parameters and 1983;63(4):1420–80. https:// doi. org/ 10. 1152/ physr ev. 1983. 63.4. 1420. disease resistance in rainbow trout (Oncorhynchus mykiss) upon synbiotic 34. Reuter SE, Evans AM. Carnitine and acylcarnitines. Clin Pharmacokinet. feeding. Fish Shellfish Immunol. 2015;45(1):27–32. https:// doi. org/ 10. 2012;51(9):553–72. https:// doi. org/ 10. 1007/ BF032 61931. 1016/j. fsi. 2015. 03. 029. 35. Ghonimy A, Zhang DM, Farouk MH, Wang Q. The impact of carnitine on 18. Hoseinifar SH, Hoseini SM, Bagheri D. Eec ff ts of galactooligosaccharide dietary fiber and gut bacteria metabolism and their mutual interaction and Pediococcus acidilactici on antioxidant defence and disease resistance in monogastrics. Int J Mol Sci. 2018;19(4):1008. https:// doi. org/ 10. 3390/ of Rainbow trout, Oncorhynchus Mykiss. Ann Anim Sci. 2017;17(1):217. ijms1 90410 08. https:// doi. org/ 10. 1515/ aoas‑ 2016‑ 0024. 36. Slotte JP. Biological functions of sphingomyelins. Prog Lipid Res. 2013;52(4):424–37. https:// doi. org/ 10. 1016/j. plipr es. 2013. 05. 001. Dhanasir i et al. Animal Microbiome (2023) 5:10 Page 21 of 22 37. Maceyka M, Spiegel S. Sphingolipid metabolites in inflammatory disease. 57. Vandeputte D, Kathagen G, D’hoe K, Vieira‑Silva S, Valles‑ Colomer M, Nature. 2014;510(7503):58–67. https:// doi. org/ 10. 1038/ natur e13475. Sabino J, et al. Quantitative microbiome profiling links gut community 38. Rolando M, Buchrieser C. A Comprehensive review on the manipulation variation to microbial load. Nature. 2017;551(7681):507–11. https:// doi. of the sphingolipid pathway by pathogenic bacteria. Front Cell Dev Biol. org/ 10. 1038/ natur e24460. 2019;7:168. https:// doi. org/ 10. 3389/ fcell. 2019. 00168. 58. Rasmussen R. Quantification on the LIGHTCYCLER. In: Meuer S, C. Wittwer 39. den Besten G, van Eunen K, Groen AK, Venema K, Reijngoud D‑ J, Bakker C, Nakagawara KI, editors. rapid cycle real‑time PCR. Berlin: Springer; 2001. BM. The role of short‑ chain fatty acids in the interplay between diet, gut 59. Hellemans J, Mortier G, De Paepe A, Speleman F, Vandesompele J. qBase microbiota, and host energy metabolism. J Lipid Res. 2013;54(9):2325–40. relative quantification framework and software for management and https:// doi. org/ 10. 1194/ jlr. R0360 12. automated analysis of real‑time quantitative PCR data. Genome Biol. 40. McLoughlin RF, Berthon BS, Jensen ME, Baines KJ, Wood LG. Short‑ chain 2007;8(2):R19. https:// doi. org/ 10. 1186/ gb‑ 2007‑8‑ 2‑ r19. fatty acids, prebiotics, synbiotics, and systemic inflammation: a systematic 60. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al‑ Ghalith GA, review and meta‑analysis. Am J Clin Nutr. 2017;106(3):930–45. https:// doi. et al. Reproducible, interactive, scalable and extensible microbiome data org/ 10. 3945/ ajcn. 117. 156265. science using QIIME 2. Nat Biotechnol. 2019;37(8):852–7. https:// doi. org/ 41. Allameh SK, Ringø E, Yusoff FM, Daud HM, Ideris A. Dietary supplement of 10. 1038/ s41587‑ 019‑ 0209‑9. Enterococcus faecalis on digestive enzyme activities, short‑ chain fatty acid 61. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello production, immune system response and disease resistance of Javanese EK, et al. QIIME allows analysis of high‑throughput community sequenc‑ carp (Puntius gonionotus, Bleeker 1850). Aquac Nutr. 2017;23(2):331–8. ing data. Nat Methods. 2010;7(5):335–6. https:// doi. org/ 10. 1038/ nmeth.f. https:// doi. org/ 10. 1111/ anu. 12397. 303. 42. Asaduzzaman M, Iehata S, Akter S, Kader MA, Ghosh SK, Khan MNA, 62. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, Holmes SP. et al. Eec ff ts of host gut ‑ derived probiotic bacteria on gut morphology, DADA2: high‑resolution sample inference from Illumina amplicon data. microbiota composition and volatile short chain fatty acids production of Nat Methods. 2016;13(7):581–3. https:// doi. org/ 10. 1038/ nmeth. 3869. Malaysian Mahseer Tor tambroides. Aquac Rep. 2018;9:53–61. https:// doi. 63. Bokulich NA, Kaehler BD, Rideout JR, Dillon M, Bolyen E, Knight R, et al. org/ 10. 1016/j. aqrep. 2017. 12. 003. Optimizing taxonomic classification of marker ‑ gene amplicon sequences 43. Macfarlane S, Macfarlane GT. Regulation of short‑ chain fatty acid produc‑ with QIIME 2’s q2‑feature ‑ classifier plugin. Microbiome. 2018;6(1):90. tion. Proc Nutr Soc. 2003;62(1):67–72. https:// doi. org/ 10. 1079/ PNS20 https:// doi. org/ 10. 1186/ s40168‑ 018‑ 0470‑z. 02207. 64. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA 44. Zou J, Secombes CJ. The function of fish cytokines. Biology. 2016;5(2):23. ribosomal RNA gene database project: improved data processing and https:// doi. org/ 10. 3390/ biolo gy502 0023. web‑based tools. Nucleic Acids Res. 2013;41(Database issue):D590–6. 45. Bird S, Tafalla C. Teleost chemokines and their receptors. Biology. https:// doi. org/ 10. 1093/ nar/ gks12 19. 2015;4(4):756–84. https:// doi. org/ 10. 3390/ biolo gy404 0756. 65. Davis NM, Proctor DM, Holmes SP, Relman DA, Callahan BJ. Simple statisti‑ 46. Shan S, Liu D, Liu R, Zhu Y, Li T, Zhang F, et al. Non‑mammalian Toll‑like cal identification and removal of contaminant sequences in marker ‑ gene receptor 18 ( Tlr18) recognizes bacterial pathogens in common carp and metagenomics data. Microbiome. 2018;6(1):226. https:// doi. org/ 10. (Cyprinus carpio L.): indications for a role of participation in the NF‑κB 1186/ s40168‑ 018‑ 0605‑2. signaling pathway. Fish Shellfish Immunol. 2018;72:187–98. https:// doi. 66. Dhariwal A, Chong J, Habib S, King IL, Agellon LB, Xia J. MicrobiomeAna‑ org/ 10. 1016/j. fsi. 2017. 09. 081. lyst: a web‑based tool for comprehensive statistical, visual and meta‑ 47. van der Vliet A, Danyal K, Heppner DE. Dual oxidase: a novel therapeutic analysis of microbiome data. Nucleic Acids Res. 2017;45( W1):W180–8. target in allergic disease. Br J Pharmacol. 2018;175(9):1401–18. https:// doi. https:// doi. org/ 10. 1093/ nar/ gkx295. org/ 10. 1111/ bph. 14158. 67. Chong J, Liu P, Zhou G, Xia J. Using MicrobiomeAnalyst for comprehen‑ 48. Bedard K, Krause K‑H. The NOX family of ROS‑ generating NADPH oxi‑ sive statistical, functional, and meta‑analysis of microbiome data. Nat dases: physiology and pathophysiology. Physiol Rev. 2007;87(1):245–313. Protoc. 2020;15(3):799–821. https:// doi. org/ 10. 1038/ s41596‑ 019‑ 0264‑1. https:// doi. org/ 10. 1152/ physr ev. 00044. 2005. 68. Martin M. Cutadapt removes adapter sequences from high‑throughput 49. Semova I, Carten JD, Stombaugh J, Mackey LC, Knight R, Farber SA, et al. sequencing reads. EMBnet J. 2011;17(1):10. https:// doi. org/ 10. 14806/ ej. Microbiota regulate intestinal absorption and metabolism of fatty acids 17.1. 200. in the zebrafish. Cell Host Microbe. 2012;12(3):277–88. https:// doi. org/ 10. 69. Kim D, Paggi JM, Park C, Bennett C, Salzberg SL. Graph‑based genome 1016/j. chom. 2012. 08. 003. alignment and genotyping with HISAT2 and HISAT‑ genotype. Nat Bio‑ 50. Falcinelli S, Picchietti S, Rodiles A, Cossignani L, Merrifield DL, Taddei AR, technol. 2019;37(8):907–15. https:// doi. org/ 10. 1038/ s41587‑ 019‑ 0201‑4. et al. Lactobacillus rhamnosus lowers zebrafish lipid content by changing 70. Tekle KM, Gundersen S, Klepper K, Bongo LA. Norwegian e‑infrastructure gut microbiota and host transcription of genes involved in lipid metabo‑ for life sciences (NeLS). F1000Research. 2018;7(1):968. https:// doi. org/ 10. lism. Sci Rep. 2015;5:9336. https:// doi. org/ 10. 1038/ srep0 9336.12688/ f1000 resea rch. 15119.1. 51. Lin R, Liu W, Piao M, Zhu H. A review of the relationship between the gut 71. Anders S, Pyl PT, Huber W. HTSeq—a Python framework to work with microbiota and amino acid metabolism. Amino Acids. 2017;49(12):2083– high‑throughput sequencing data. Bioinformatics. 2014;31(2):166–9. 90. https:// doi. org/ 10. 1007/ s00726‑ 017‑ 2493‑3.https:// doi. org/ 10. 1093/ bioin forma tics/ btu638. 52. Rowland I, Gibson G, Heinken A, Scott K, Swann J, Thiele I, et al. 72. Anders S, Huber W. Differential expression analysis for sequence Gut microbiota functions: metabolism of nutrients and other food count data. Genome Biol. 2010;11(10):1–12. https:// doi. org/ 10. 1186/ components. Eur J Nutr. 2018;57(1):1–24. https:// doi. org/ 10. 1007/ gb‑ 2010‑ 11‑ 10‑ r106. s00394‑ 017‑ 1445‑8. 73. Reimand J, Kull M, Peterson H, Hansen J, Vilo J. g:Profiler–a web ‑based 53. FEEDAP. Assessment of the application for renewal of authorisation of toolset for functional profiling of gene lists from large ‑scale experiments. Bactocell (Pediococcus acidilactici CNCM I‑4622) as a feed additive for Nucleic Acids Res. 2007;35( Web Server issue):W193–200. https:// doi. org/ all fish and shrimps and its extension of use for all crustaceans. EFSA J. 10. 1093/ nar/ gkm226. 2019;17(4):e05691. https:// doi. org/ 10. 2903/j. efsa. 2019. 5691. 74. Raudvere U, Kolberg L, Kuzmin I, Arak T, Adler P, Peterson H, et al. 54. Thorarensen H, Farrell AP. The biological requirements for post‑ g:Profiler: a web server for functional enrichment analysis and conver ‑ smolt Atlantic salmon in closed‑ containment systems. Aquaculture. sions of gene lists (2019 update). Nucleic Acids Res. 2019;47( W1):W191–8. 2011;312(1):1–14. https:// doi. org/ 10. 1016/j. aquac ulture. 2010. 11. 043.https:// doi. org/ 10. 1093/ nar/ gkz369. 55. Knudsen BE, Bergmark L, Munk P, Lukjancenko O, Priemé A, Aarestrup FM, 75. Supek F, Bošnjak M, Škunca N, Šmuc T. REVIGO summarizes and visualizes et al. Impact of sample type and DNA isolation procedure on genomic long lists of gene ontology terms. PLoS ONE. 2011;6(7):e21800. https:// inference of microbiome composition. mSystems. 2016;1(5):e00095. doi. org/ 10. 1371/ journ al. pone. 00218 00. https:// doi. org/ 10. 1128/ mSyst ems. 00095‑ 16. 76. Evans AM, DeHaven CD, Barrett T, Mitchell M, Milgram E. Integrated, 56. Ramseier CA, Kinney JS, Herr AE, Braun T, Sugai JV, Shelburne CA, et al. nontargeted ultrahigh performance liquid chromatography/electrospray Identification of pathogen and host ‑response markers correlated with ionization tandem mass spectrometry platform for the identification and periodontal disease. J Periodontol. 2009;80(3):436–46. https:// doi. org/ 10. relative quantification of the small‑molecule complement of biological 1902/ jop. 2009. 080480. Dhanasiri et al. Animal Microbiome (2023) 5:10 Page 22 of 22 systems. Anal Chem. 2009;81(16):6656–67. https:// doi. org/ 10. 1021/ ac901 536h. 77. Evans AM, Br B, Liu Q, Mitchell MW, Rj R, Dai H, et al. High resolution mass spectrometry improves data quantity and quality as compared to unit mass resolution mass spectrometry in high‑throughput profiling metabolomics. Metabolomics. 2014;4:1–3. 78. Ni Y, Yu G, Chen H, Deng Y, Wells PM, Steves CJ, et al. M2IA: a web server for microbiome and metabolome integrative analysis. Bioinformatics. 2020;36(11):3493–8. https:// doi. org/ 10. 1093/ bioin forma tics/ btaa1 88. 79. You Y, Liang D, Wei R, Li M, Li Y, Wang J, et al. Evaluation of metabolite– microbe correlation detection methods. Anal Biochem. 2019;567:106–11. https:// doi. org/ 10. 1016/j. ab. 2018. 12. 008. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions
Journal
Animal Microbiome – Springer Journals
Published: Feb 11, 2023
Keywords: Functional ingredients; Prebiotics; Probiotics; FOS; GOS; Gut microbiota; Pediococcus acidilactici; Metabolomics; Transcriptomics; Atlantic salmon