Summarzing Mining and Bioactivity Results from Publicly Available Fermented Foods Genomes and Peptidomics Data

This repository contains scripts and notebooks for analyzing data produced by the bac-mining workflow and peptide-bioactivity-prediction workflow for the following datasets:

1. 5 Peptidomics Studies from Fermented Foods
2. ~200 Bacterial Isolates from the BacDive database collected from various fermented foods
3. ~11,500 bacterial metagenome-assembled genomes (MAGs) from diverse fermented food metagenomic surveys

The fermentefood_mags_curation repository documents how all of the MAG and BacDive isolate genome data were collected and processed.

Repository Structure

This repository is setup so that all processed results from the workflows are in results, they are analyzed and viewed within the notebooks in the notebooks directory, which call functions from the scripts/mag-mining-notebook-functions.R script. The subdirectory structure is as follows:

figures/ - Figures output from notebooks and figure generation scripts
metadata/ - Genome and proteomics sample metadata, most of the metadata is pulled from a controlled repository link with the most up-to-date genome metadata version
notebooks/ - Analysis notebooks
raw_data/ - Raw data for batches of bac-mining runs, BacDive genomes FASTA file, and raw FASTA files for the proteomics experiments
results/ - Results split up by dataset for outputs from the bac-mining and peptide-bioactivity-predictor workflows, as well as cleaned results
scripts/ - Scripts for notebook functions, figure generation, parsing and combining FASTA files for feeding into workflows for batch runs, and parsing raw files for the proteomics experiments

## Results

Raw files including FASTA sequences, machine learning models, and peptides results for all three data sources are [available on Zenodo](https://zenodo.org/records/16749254).