iHOP

iHOP provides literature-derived, sentence-level linkage of genes and proteins extracted from PubMed abstracts to enable exploration and analysis of gene and protein interactions.

Key Features:

• Automated sentence extraction: Automatically extracts key sentences from millions of PubMed abstracts to summarize literature for biological entities.
• Gene/protein hyperlinking and literature network: Converts gene and protein mentions in abstracts into hyperlinks to create an interconnected, sentence-level literature network.
• Coverage: Contains approximately five million sentences and information on about 40,000 genes and over 80,000 biological entities across Homo sapiens, Mus musculus, Drosophila melanogaster, Caenorhabditis elegans, Danio rerio, Arabidopsis thaliana, Saccharomyces cerevisiae, and Escherichia coli.
• Programmatic API: Provides a public programmatic API exposing core functionalities for integration into bioinformatics programs and workflows.
• Data integration: Supports superimposing experimental interaction data onto the literature network for simultaneous analysis of novel and published interactions.

Scientific Applications:

• Literature-based interaction discovery: Enables identification and exploration of gene and protein interactions supported by sentence-level evidence from PubMed abstracts.
• Entity-centric literature summarization: Provides concise, sentence-level summaries for genes, proteins, and other biological entities to survey published knowledge.
• Experimental data contextualization: Facilitates validating and interpreting experimental interaction datasets by overlaying them onto the literature network.
• Cross-species comparison: Supports comparative analysis of gene and protein information across multiple model organisms listed in the coverage.

Methodology:

Automatically extracts key sentences from PubMed abstracts, links gene and protein mentions as hyperlinks to build an interconnected literature network, exposes functionality via a public programmatic API, and permits superimposition of experimental interaction data onto that network.