Construction of a dairy microbial genome catalog opens new perspectives for the metagenomic analysis of dairy fermented products

Abstract

Background: Microbial communities of traditional cheeses are complex and insufficiently characterized. The origin, safety and functional role in cheese making of these microbial communities are still not well understood. Metagenomic analysis of these communities by high throughput shotgun sequencing is a promising approach to characterize their genomic and functional profiles. Such analyses, however, critically depend on the availability of appropriate reference genome databases against which the sequencing reads can be aligned.

Results: We built a reference genome catalog suitable for short read metagenomic analysis using a low-cost sequencing strategy. We selected 142 bacteria isolated from dairy products belonging to 137 different species and 67 genera, and succeeded to reconstruct the draft genome of 117 of them at a standard or high quality level, including isolates from the genera Kluyvera, Luteococcus and Marinilactibacillus, still missing from public database. To demonstrate the potential of this catalog, we analysed the microbial composition of the surface of two smear cheeses and one blue-veined cheese, and showed that a significant part of the microbiota of these traditional cheeses was composed of microorganisms newly sequenced in our study.

Conclusions: Our study provides data, which combined with publicly available genome references, represents the most expansive catalog to date of cheese-associated bacteria. Using this extended dairy catalog, we revealed the presence in traditional cheese of dominant microorganisms not deliberately inoculated, mainly Gram-negative genera such as Pseudoalteromonas haloplanktis or Psychrobacter immobilis, that may contribute to the characteristics of cheese produced through traditional methods.

Figure 1

Origin of the 142 selected dairy bacterial isolates in function of the type of dairy product (A) and the geographic area (B).

Figure 2

Global phylogeny of the 117 dairy bacterial isolates sequenced in the present study. The phylogenetic tree is an ITOL circular visualization [68] with the branch length and the bootstrap values displayed. The tree is based on a concatenated alignment of 40 universal marker protein families [40]. Only genome sequences from which a minimum of 10 markers could be extracted and which had no contaminating sequences evidence were considered. The genome of Methanobrevibacter smithii ATCC35061 was used to root the tree. The colors correspond to the different phyla.

Figure 3

Mapping of the good quality reads from the metagenomic sequencing of DNA from the surfaces of three cheeses. The good quality reads coming from 3 samples of cheese surface were aligned to 5873 genomes coming from NCBI and 117 genomes coming from our project. The repartition of the good quality reads that map only on the NCBI genomes (blue), on the genome sequenced in our project (green), on both NCBI and our genome (light green) and on Bos taurus genome (orange) is presented in pie charts. The unmapped good reads are presented in dark and light grey, respectively those lacking a reference and those potentially unmappable for technical reason.