PANTHER : Protein families and subfamilies modeled on the divergence of function

Abstract

Protein Analysis Through Evolutionary Relationships PANTHER is a resource that defines a molecular taxonomy of proteins, and relates this taxonomy to computational representations of function, spanning molecular functions, biological processes, and biochemical pathways. In this way, PANTHER provides an infrastructure for exploring the relationships between protein sequence evolution and evolution of function at both the molecular and pathway levels. Proteins are grouped into families, and the relationships between sequences in each family are represented as a phylogenetic tree. The trees are constructed using sequence conservation‐based distances that may reflect the functional constraints, over evolutionary time, on proteins of the same or similar functions. The current version of PANTHER contains calculated trees for over 5000 protein families, spanning over 80% of the protein‐coding genes in mammalian genomes and their known relatives from all kingdoms of life. Expert biologists have divided these families into over 30 000 subfamilies (clades), where each subfamily comprises proteins that share a distinct function. Subfamilies therefore represent the functional divergence (or “cooption”) events that occurred during the family's evolutionary history. Each subfamily is associated, when possible, with ontology terms that describe the functions of its constituent proteins and, increasingly, with components in biochemical pathways. In addition, the sequences in each family and subfamily are represented as a statistical amino acid “signature”, allowing classification of new sequences, and exploration of the relationships between protein sequence and function. PANTHER is freely available at http://www.pantherdb.org , where users can submit protein sequences for classifications, search, and browse the classifications and trees, and employ computational tools to analyze their own data, such as gene expression data and genetic mutation data, in the context of protein families and functions.

PANTHER version 10: expanded protein families and functions, and analysis tools

“…Protein sequences (one representative sequence per gene) from 104 complete genomes (PANTHER uses the UniProt Reference Proteomes data set ( 1 )) are organized into families of homologous genes. For each family, a phylogenetic tree is constructed, representing the evolutionary relationships between all genes in the family, and the processes by which these genes were first copied and then diverged from each other: speciation, gene duplication and horizontal transfer ( 2 , 3 ). In speciation, a gene is copied when an ancestral population separates into two or more isolated populations.…”