Historically, microbial type strains are defined as the descendants of original isolates employed in the first description of a novel taxa, commonly resulting in their extensive use as model or prototypical strains. Nevertheless, since identification of such type strains would generally have occurred before the advent of the microbial genomics era, they may not correspond to the most representative of their species from an ecological, genomic and functional perspective.

In the current study, we revised the concept of type strain based on a comprehensive multi-omics approach using the Bifidobacterium genus as an example and abundant microbial taxon of the human Infant gut microbiota. Strain tracking based on 1664 publicly available shotgun metagenomics datasets of healthy infant fecal samples were employed together with genomics investigations and screening of international public biobanks to identify the most relevant reference bifidobacterial strains, by being a representative of their whole species in the infant population, and suitable for in silico and in vitro analyses. Subsequently, an ad hoc bioinformatic tool was developed allowing the screening of local strain collections to identify the most suitable strain as a species representative based on phylogenetic relationship to the corresponding proposed model species.

The here proposed approach for the identification of novel reference strains was further validated using in vitro trials followed by metagenomics and metatranscriptomics analyses. This was performed by simulating both the microbe-microbe and microbe-host interactions of the proposed reference strains through cultivation in a bioreactor gut model in the presence of a human gut microbiota as well as by their close contact with the human cell lines monolayers. Altogether, these results demonstrated the validity of the model strain selection based on ecological and genomic data, resulting in improved in silico and in vitro investigations both in terms of cross-laboratory reproducibility and broadness of findings.