Strain-level resolution for starter culture monitoring of cocoa fermentation processes. CR2019

A wide range of microbiological techniques have been applied to monitor the growth and prevalence of starter culture strains used in the production of fermented foods. Differentiating between inoculated starter culture strains and background strains from the same species spontaneously present in the fermenting matrices has been challenging. However, the intra-species diversity present in the microbiota of fermented food ecosystems plays an important role in the strain fitness for fermentation in terms of substrate utilization and stress resistance and is of major importance for their contribution to the formation of flavour compounds. The aim of the current study was to monitor starter culture strains inoculated to steer cocoa fermentation processes performed in Costa Rica at strain-level resolution. First, whole-community DNA-based, high-throughput, full-length 16S rRNA gene sequencing was performed using PacBio circular consensus sequencing and amplicon sequence variants (ASVs) were inferred using DADA2. Further, three shotgun metagenomics-based approaches were followed, the first relying on species-specific marker genes using StrainPhlAn, a second one relying on single nucleotide variants (SNVs) retrieved from the metagenomic reads, and a third one relying on metagenome-assembled genome (MAG) reconstruction followed by average nucleotide identity (ANI) estimation. Albeit that the ASV-based strategy was time- and cost-effective to study microbial community dynamics at strain-level resolution, the application of novel metagenomic strategies provided more information and slightly higher resolution to recover the inoculated starter culture strains from complex ecosystems. Further, resolving different strains of the same species into independent MAGs allowed the study of intra-species differences at gene level that could be of interest to guide future starter culture trials for any fermented food ecosystem.