The Development of a Functional Annotation Pipeline to Characterise Metagenome-Assembled Genomes of Microorganisms Found in Anaerobic Digestion

Anaerobic digestion involves the conversion of organic waste into biogas and biofertilisers. Anaerobic digesters are commonly found within the wastewater treatment process in the UK, converting waste sludge into methane. Higher yields of methane are required for AD to become a favourable renewable energy source. The AD process consists of four steps (hydrolysis, acidogenesis, acetogenesis, and methanogenesis) that are driven by complex microbial communities. Hydrogenotrophs and methanogens are rate-determining factors, highlighting the significance of these microbial communities within these dynamic AD environments. Research into these microbial communities will ultimately result in greater yields of methane in AD.

A greater understanding of the microbial communities can be achieved via metagenomics, which involves the study of genomes recovered from environmental samples. Metagenomics involves the use of shotgun sequencing. Environmental DNA is sequenced followed by binning, and assembly into metagenome-assembled genomes (MAGs). Functional annotation is carried out to predict the gene function within the MAGs. However, quality and completeness of MAGs varies greatly due to the nature of shotgun sequencing. Large datasets of metagenomic data require large-scale data manipulation and bioinformatic analysis. Genome annotation pipelines (via workflow management tools e.g. Snakemake) allow automation and ensure reproducibility of the genome annotation.

A genome annotation pipeline was developed, using Snakmake, to predict the gene function of MAGs recovered from AD. This pipeline was developed to provide an automated tool to functionally annotate MAGs, in order to discover more about the metabolic processes and relationships between microbes that drive the AD process. A confidence system was devised to indicate the quality of annotations provided by orthology-based tools EggNOG and KofamScan, allowing further analysis of low quality ORFs. Reproducibility and reference databases continue to be limitations of bioinformatic pipelines. However, approximately 50% of ORFs are annotated to a high confidence.