Protocols for Representation of Protein Domain Annotations in Clade-Oriented Databases: a Case Study at the Legume Information System using Chado/Tripal

Protein domains can be conceptualized from a number of perspectives, from their role in defining an individual protein’s structure and function to their evolutionary role in creating novel molecular functions through duplication and recombination into unique multi-domain protein architectures. Although many species- and clade-oriented databases use standard protein domain analyses to characterize the putative functions and cellular localizations of the gene products represented in the genomes and transcriptomes of their species of interest, this is often limited to treating the matched domains as properties of the genes that are simply an aid to their classification and retrieval. While this gene-centric perspective is clearly of great importance, elevating domains to a prominent position in the context of such databases has the potential to provide insights into many interesting biological questions, from the role of domains in constraining and shaping intra-species diversity patterns (including SNPs, splice isoforms, and gene fusions) to their role in providing the basis for the definition of gene family groupings of orthologous and paralogous genes as well as providing insights into their evolutionary dynamics. We have utilized and extended a set of widely used open source tools for analysis, storage and web-based presentation of protein domain data to populate the Chado database underlying the Legume Information System (http://legumeinfo.org) and to make this data available through a shareable Tripal extension module for enabling search by domains, exploiting the ontological structure of InterPro and interlinking our domain-oriented representation to other modules for presentation of gene and gene families.