IggyTop: a unified prior knowledge resource for immunoreceptor-epitope interactions with automated data integration

Immune system diversity has been extensively studied for years, generating vast amounts of data distributed across numerous repositories. A significant portion of this research focuses on the specificity of different immunoreceptors. This specificity information serves as valuable prior knowledge for analyzing patient immunoreceptor sequencing data, where matching epitopes can provide insights into the pathogens affecting the patient or serve as potential treatment targets. The larger the prior knowledge pool, the higher the probability of finding a matching epitope for a given sequence. Several existing resources collect immunoreceptor-epitope interaction data; however, they have distinct limitations: different research focuses (e.g., cancer versus pathogens), targeting of specific immunoreceptor types (e.g., T cell receptors only), and reliance on manual curation and updates. To increase coverage in receptor-epitope matching, we developed IggyTop, a unified resource containing immunoreceptor-epitope pairing combinations from existing databases: IEDB, VDJdb, McPAS-TCR, CEDAR, TRAIT, NeoTCR, and TCR3d. IggyTop is open source, and the modular architecture of the underlying BioCypher framework enables seamless integration of additional databases without requiring modifications to existing components, providing flexibility for future expansion. The resource is automatically rebuilt regularly and harmonizes data on-the-fly, ensuring that updates from different databases are automatically integrated into the resource. IggyTop is stored as a JSON file containing a list of AIRR cells as defined in scirpy, where each cell is an awkward array holding information about chain(s) and matching epitopes.