HPIDB: A Database for Annotation, Computational Prediction and Analysis of Host-Pathogen Interactions

Infectious disease researchers use host-pathogen interactions (HPI) to understand the complex interplay between host and pathogen, and to identify novel therapeutic and intervention strategies. The Host-Pathogen Interaction Database (HPIDB) is a public resource that facilitates modeling and analysis of HPIs. This resource integrates existing HPI information and biocurators provide manually curated annotations for animal. We will present our current progress and future annotation goals, focusing on host-pathogen pairs that we will curate; we are also requesting feedback from the community about these curation targets. At present, HPIDB contains 52,953 curated entries which include a broad range of species (58 hosts and 524 pathogens). The downloadable HPIDB data can be combined with other HPIDB services such as network visualization and ontology functional information to better inform our understanding of infectious diseases. In addition to manually annotated HPIs, HPIDB serves as a platform to infer additional HPI and support investigator-driven data analysis where annotated data is scarce. For an improved computational HPI prediction, we are developing a prediction model by comparing multiple protein-related features, algorithms, and datasets for prediction of HPIs. We expect to develop this workflow as an online tool so that researchers can rapidly predict and model interactions for their own pathogens of interest; we expect that this modeling will complement wetlab work to identify potential targets. As we develop the HPI prediction tool we will work closely with researchers who have their own animal health data sets to ensure that this resource meets the community needs.