The microbiome’s fiber degradation profile and its relationship with the host diet

Abstract

Background The relationship between the gut microbiome and diet has been the focus of numerous recent studies. Such studies aim to characterize the impact of diet on the composition of the microbiome, as well as the microbiome’s ability to utilize various compounds in the diet and produce metabolites that may be beneficial for the host. Consumption of dietary fibers (DFs)—polysaccharides that cannot be broken down by the host’s endogenous enzymes and are degraded primarily by members of the microbiome—is known to have a profound effect on the microbiome. Yet, a comprehensive characterization of microbiome compositional and functional shifts in response to the consumption of specific DFs is still lacking. Results Here, we introduce a computational framework, coupling metagenomic sequencing with careful annotation of polysaccharide degrading enzymes and DF structures, for inferring the metabolic ability of a given microbiome sample to utilize a broad catalog of DFs. We demonstrate that the inferred fiber degradation profile (IFDP) generated by our framework accurately reflects the dietary habits of various hosts across four independent datasets. We further demonstrate that IFDPs are more tightly linked to the host diet than commonly used taxonomic and functional microbiome-based profiles. Finally, applying our framework to a set of ~700 metagenomes that represents large human population cohorts from 9 different countries, we highlight intriguing global patterns linking DF consumption habits with microbiome capacities. Conclusions Combined, our findings serve as a proof-of-concept for the use of DF-specific analysis for providing important complementary information for better understanding the relationship between dietary habits and the gut microbiome.

Mapping gut bacteria into functional niches reveals the ecological structure of human gut microbiomes

“…Our approach complements other methods to describe how gut microbiome function is structured 42 . Importantly, it provides much needed quantitative biological data to inform computational models 43 .…”

Identification of carbohydrate gene clusters obtained from in vitro fermentations as predictive biomarkers of prebiotic responses

“…Given the vast diversity of substrates available, finer CAZyme subfamily resolutions will be necessary to aid the discovery of other prebiotic associated genes to allow for a greater scale of precision. Improvements in CAZyme annotation and curation frameworks as described in Cohen & Borenstein [61]will be especially beneficial in capturing fiber degradative profiles and subsequent biomarker discovery across fiber intervention studies.…”

“…Despite evidence of functional redundancy in the gut microbiome [62,63], shifts in CAZyme composition have previously been observed in response to different substrate feeding [64,65]. Likewise, in population-based studies, CAZyme profiles have also been shown to reflect dietary preferences and the type of glycans that microbes encounter in the gut [61,66,67]. Moreover, strain heterogeneity in CAZyme profiles within the same genus has been described such as in species of Bifidobacterium and Prevotella [68,69].…”

Carbohydrate-active enzyme annotation in microbiomes using dbCAN

“…1), featuring glycan substrate prediction for both CAZymes and CGCs (Table 1). This substrate prediction function is critical to microbiome researchers who are interested in knowing not only what CAZyme families exist in their genomes or metagenomes, but also what glycan substrates can be metabolized in their samples (e.g., by bacterial isolates or microbiota) [42][43][44] . For example, a paper published in 2018 compared human gut microbiomes of two groups of patients with type 2 diabetes 45 .…”

“…In addition to assembly-based approaches, the assembly-free strategy is also used, e.g., in the popular general metagenomic functional annotation tool HUMAnN3 57 and the inferred fiber degradation profile tool (IFDP) 43 . The assembly-free approach does not need run_dbcan but rather directly searches sequencing reads against the CAZy annotated protein sequence database (i.e., CAZyDB).…”