PANTHER Tools

PANTHER Tools classifies protein sequences into families and subfamilies and associates them with functional ontologies and pathways to support large-scale functional, evolutionary, and SNP-impact analyses.

Key Features:

• Comprehensive Protein Family Database: As of version 5.0 PANTHER comprises 6,683 protein families divided into 31,705 subfamilies, covering approximately 90% of mammalian protein-coding genes.
• Functional Divergence Modeling: Models functional divergence within protein families by associating families and subfamilies with ontology terms and pathways and identifying amino acids crucial for functional specificity.
• Hidden Markov Models (HMMs): Employs HMMs built for each family and subfamily to classify additional protein sequences and predict function.
• Pathway Representation: Represents pathways, including signaling pathways, associated with subfamilies and individual protein sequences.
• Scoring and Analysis Tools: Provides resources to score sequences against PANTHER HMMs and analyze large gene lists from high-throughput expression experiments.
• Ontology and Indexing: Implements a simplified protein function ontology and the PANTHER index (PANTHER/X) summarizing molecular functions and biological processes associated with families and subfamilies.

Scientific Applications:

• Functional Characterization: Relates sequence relationships to function relationships for large-scale protein function annotation.
• Gene Expression Analysis: Supports analysis of high-throughput expression data by scoring proteins against the HMM library and mapping results to functions and pathways.
• SNP Impact Assessment: Ranks missense single nucleotide polymorphisms by their likelihood of affecting protein function to inform studies of Mendelian and complex diseases.
• Evolutionary Analysis: Uses evolutionary comparisons to assess molecular effects of disease-associated SNPs across different disease types.
• Integration with Other Databases: Integrates and coordinates annotations with external resources such as InterPro and PIRSF.

Methodology:

Uses Hidden Markov Models built per family and subfamily to classify sequences; models functional divergence by associating families/subfamilies with ontology terms and pathways and identifying specificity-determining amino acids; scores sequences against HMMs; and performs evolutionary analyses to compare disease-associated SNPs and rank missense variants by predicted functional impact.