The PANTHER System is constructed with three functional modules (Figure 1). The core module is a large protein library that contains all protein-coding genes from 82 organisms organized first into families based on sequence homology, and then into subfamilies based on their shared functions (often a group of orthologous genes). Each family or subfamily is represented by a statistical model (HMM) annotated with ontology terms (GO terms and PANTHER Protein Class terms). The second module is the PANTHER pathway module, which contains 176 expert curated pathways. All pathways are connected, through manual curation, to individual proteins in the protein library through the pathway components, and therefore they are also linked to the phylogenetic information and statistical models. The last module is the website tool suite that contains a collection of bioinformatics tools and software, enabling users to not only query the data and classify genes and proteins, but also visualize, analyze and interpret genome-wide scale experimental data in the context of the enriched data content in the first two modules.
