Metagenomic binning

The contigs from both co-assembly and single-sample assembly were filtered based on the sequence length. The contigs with sequence length > 1500 bp were retained by seqtk (https://github.com/lh3/seqtk) and used for metagenomic binning. For metagenomic binning analysis, the clean reads were mapped to corresponding contigs using Bowtie2 (v. 2.4.4) [26]. Samtools (v. 1.7.0) was used to convert mapped results into BAM format [27]. Then, the BAM files were sorted and indexed using SAMtools (v. 1.7.0) [27]. The resulting sorted BAM files were used for metagenomic binning based on the sequence characteristics and coverage depth using MaxBin2 (v. 2.2.7) [28], MetaBAT2 (v. 2:2.15) [29] and CONCOCT (v. 1.1.0) [30]. DAS Tool (v. 1.1.2) [31] was then applied to integrate MAGs generated from different methods.

The completeness and contamination of all MAGs were estimated using CheckM (v. 1.0.12) [32] based on the lineage_wf workflow. The MAGs with medium and high qualities (completeness ≥ 50% and contamination ≤ 10%) were retained. The retained MAGs were classified into 15 datasets based on food fermentation types. Then, MAGs from each food fermentation type were dereplicated using fastANI algorithm in dRep (v. 3.2.2) [33] at the threshold of 99% average nucleotide identity (ANI) (strains level) with at least 25% overlap between genomes. Meanwhile, to enhance the diversity of the dataset, the publicly available 328 MAGs in cheese fermentation [21] and 29 MAGs in cocoa fermentation [34] were compared with the MAGs in cheese and cocoa fermentations in this study, respectively. We removed the repeated MAGs between publicly available and our MAGs based on 99% ANI using dRep (v. 3.2.2) [33]. There were 27 and 18 different bacterial MAGs in cheese fermentation [21] and cocoa fermentation [34], respectively. These different MAGs were added in the corresponding food fermentation types in this study. A total of 653 bacterial MAGs were obtained from 15 food fermentation types. These 653 MAGs were nonredundant MAGs recovered after combination with public data and dereplication. All MAGs were taxonomically annotated using GTDB-Tk (v. 0.1.6) [35] based on the Genome Taxonomy Database (http://gtdb.ecogenomic.org/), and the standardised taxonomic labels were obtained. The detailed commands in metagenomic assembly and binning analysis are available at https://github.com/durubing-jn/food-fermentation-mategenome.