A shotgun metagenomics analysis applied to the ecological engineering of anaerobic microbial communities for the production of hydrogen from Citrus Peel Waste
Résumé
Brazil, the first Citrus fruit producer worldwide, produces annually 88 million tons of this fruit; a large part of it is processed, generating waste, composed mainly by peels and bagasse, which accounts for half of the processed mass. More than half of this waste is not yet valorized, although it could be used in biorefineries to produce various invaluable compounds, including biogas. In this project, we focused on the production of hydrogen from Citrus Peel Waste by anaerobic microbial communities. We used a shotgun metagenomics approach to identify key strains and biological functions for this process, by comparing the inoculum (not yet adapted) with the microbial communities sampled after hydrogen production in optimized conditions.
We developed a bioinformatics pipeline for shotgun metagenomics data analysis. After a preprocessing step, reads were co-assembled using metaSPADES. Coding regions were predicted with Prodigal. The taxonomic assignments of the predicted genes were obtained with kaiju against NCBI nr database. For the functional annotation, the predicted genes were affiliated with ghostKoala against KEGG database and with dbCAN against the CAZY database. For each dataset individually, reads were mapped on the co-asssembled contigs. It was thus possible to build a table, which included, for each predicted gene, the counts by sample as well as the functional and taxonomic annotations.
This pipeline was developed with snakemake to favour reproducible and scalable data analysis. It was run on the cluster of the INRAE MIGALE bioinformatics platform. It is accessible to the community on GitLab (https://gitlab.irstea.fr/cedric.midoux/workflow_metagenomics).
We were able to identify key strains and biological functions selected during this process, especially those related to lignocellulose deconstruction and to hydrogen production. A strong and reproducible microbial selection was observed, leading to the dominance of C. beijerinckii and C. butyricum, known to be hydrogen producers. The glycoside hydrolase 48 family, thought to be critical for cellulose hydrolysis, was in low proportion and it was associated to C. puniceum, a cellulosome-producing bacterium. C. puniceum could thus in the present case be key for some steps of the lignocellulose deconstruction, despite its low proportion.
Domaines
Bio-informatique [q-bio.QM]Origine | Fichiers produits par l'(les) auteur(s) |
---|