MetaCerberus

MetaCerberus performs high-throughput, HMM-based functional annotation of genes across single genomes and complex metacommunities to infer gene functions and metabolic pathways.

Key Features:

• HMM-based inference: Uses Hidden Markov Model (HMM) methodologies implemented via HMMER for gene function prediction.
• Massively parallelized architecture: Employs parallelization to reduce memory usage and accelerate annotation throughput.
• Performance vs. eggNOG-mapper v2: Reports significantly lower memory usage and a 1.3-fold increase in speed on a single computational node compared to eggNOG-mapper v2.
• Database integration: Integrates annotations from KEGG (KO), COGs, CAZy, FOAM, VOGs, and PHROGs.
• Viral and phage annotation: Optimized for viruses, phages, and archaeal viruses and reported to outperform DRAM, Prokka, and InterProScan in those domains.
• KO annotation efficiency: Annotates a greater number of KEGG Orthologs (KOs) than DRAM while using a database reported to be 186 times smaller and consuming 63 times less memory.
• Statistical and pathway integration: Supports automatic downstream analysis with DESeq2 and edgeR and pathway enrichment via GAGE R and pathview R.
• Implementation: Implemented in Python.

Scientific Applications:

• Genome and metacommunity annotation: Functional annotation of individual genomes and complex environmental metacommunities.
• Viral genomics: Annotation and functional characterization of viruses, phages, and archaeal viruses.
• Differential and pathway analysis: Comparative functional profiling and pathway enrichment analyses using DESeq2, edgeR, GAGE R, and pathview R.
• Resource-constrained large-scale studies: High-throughput gene function inference for large metagenomic datasets with reduced memory requirements.

Methodology:

HMM-based sequence annotation using HMMER with a massively parallelized architecture; automated downstream analyses using DESeq2, edgeR, GAGE R, and pathview R; implemented in Python.