SUPERFAMILY

SUPERFAMILY assigns protein domains at the SCOP superfamily level for genomes, metagenomes, and protein collections to support structural and functional annotation.

Key Features:

• Genomic and metagenomic coverage: Domain assignments span over 1,400 completely sequenced genomes and more than 120 metagenomes and extend to other gene collections such as UniProt.
• SCOP v1.75 HMM library: A hidden Markov model library derived from SCOP version 1.75, including classes such as coiled coils, for domain recognition.
• HMMER3 scoring: Sequence scoring and assignment use HMMER3 for faster processing and improved results.
• Domain-based GO annotations: GO annotations are provided at both superfamily and family levels using a novel methodology for high-quality assignment.
• Domain-based phenotypic ontologies: The same annotation methodology has been applied to generate domain-based phenotypic ontologies.
• Reference tree of life mapping: A refined reference tree of life enables highlighting of superfamilies, families, and domain architectures in an evolutionary context.
• Cloud-based computation: A cloud-based pipeline implemented on Amazon Web Services EC2 provides scalable computation for analyses.

Scientific Applications:

• Protein structural classification: Assigns proteins to SCOP superfamilies to support structural classification across organisms and datasets.
• Functional annotation: Provides domain-based GO term assignments to infer molecular function, biological process, and cellular component at domain levels.
• Comparative genomics and metagenomics: Enables comparative analyses of domain content across genomes and metagenomes for evolutionary and ecological studies.
• Phenotype ontology generation: Supports creation of domain-based phenotypic ontologies for linking domain presence to phenotypic traits.
• Evolutionary mapping: Facilitates mapping of domain architectures and superfamily distributions onto a reference tree of life for evolutionary inference.

Methodology:

Uses an HMM library based on SCOP v1.75 (including coiled coils) for domain models; sequence scoring with HMMER3; a novel domain-based GO annotation methodology (applied also to phenotypic ontologies); a refined reference tree of life for mapping; and a cloud-based pipeline on AWS EC2 for scalable computation.