Functional Profiling

A manually curated database of orthologous gene groups that represent functional units across all domains of life. Each KO entry (K-number) corresponds to a conserved functional node in the KEGG pathway maps. Functional profiling tools like HUMAnN and PICRUSt map predicted genes to KO identifiers to reconstruct community-level metabolic potential. The hierarchical structure links genes to pathways, modules, and BRITE functional hierarchies, enabling multi-resolution functional analysis.

The evolutionary genealogy of genes: Non-supervised Orthologous Groups database provides hierarchical orthology assignments across 5,000+ organisms. Unlike KO, eggNOG uses unsupervised clustering algorithms to group genes into Orthologous Groups (OGs) at multiple taxonomic levels. Each OG is annotated with functional descriptors, COG categories, and Gene Ontology terms. Tools like eggNOG-mapper use precomputed HMM profiles for rapid functional annotation of metagenomic contigs.

Pfam is a database of protein families defined by multiple sequence alignments and hidden Markov model (HMM) profiles. Each family represents a functional domain or conserved region. InterPro integrates Pfam with other signature databases (PROSITE, SMART, CDD) to provide a unified classification system. Functional profiling pipelines use HMMER to scan predicted metagenomic peptides against Pfam, identifying conserved domains that indicate enzymatic function even for divergent sequences.

MetaCyc is a highly curated, non-redundant database of experimentally elucidated metabolic pathways from all domains of life. Unlike KEGG, MetaCyc focuses exclusively on pathways with direct experimental evidence. The associated Pathway Tools software suite enables prediction of metabolic networks from annotated genomes. In metagenomic functional profiling, MetaCyc pathways provide higher-confidence functional assignments because each pathway is empirically validated rather than computationally inferred.

The Carbohydrate-Active enZYmes database classifies enzymes that build and break down complex carbohydrates and glycoconjugates. Organized into families based on sequence and structural similarity, CAZy covers: Glycoside Hydrolases (GH), Glycosyltransferases (GT), Polysaccharide Lyases (PL), Carbohydrate Esterases (CE), and Auxiliary Activities (AA). Specialized tools like dbCAN2 use HMM profiles to annotate CAZyme families in metagenomic data, critical for gut microbiome and biomass degradation studies.

The Comprehensive Antibiotic Resistance Database uses a structured ontology—the Antibiotic Resistance Ontology (ARO)—to organize resistance determinants by mechanism, drug class, and genetic context. For functional profiling, CARD's Resistance Gene Identifier (RGI) maps metagenomic reads against curated reference sequences and detects SNP-based resistance variants. This enables not just presence/absence calls but mechanism-level characterization of the resistome, distinguishing efflux pumps from target modification enzymes.

A within-sample normalization metric essential for comparing functional gene abundances across samples of different sequencing depths. RPKM corrects for gene length bias and library size by dividing the number of reads mapped to a gene by the gene's length in kilobases and the total number of mapped reads in millions. This normalization is a prerequisite for accurate differential abundance testing in functional metagenomics.