ProPheno

ProPheno aggregates mentions of human proteins and phenotypes from the complete corpora of Medline and PubMed Central Open Access to characterize protein-phenotype relationships for the study of rare and complex diseases.

Key Features:

• Extensive dataset: Constructed from the complete corpora of Medline and PubMed Central Open Access, aggregating mentions of human proteins and phenotypes.
• Textual co-occurrences: Captures co-occurrences of protein-phenotype pairs across textual spans, including sentences and paragraphs, to preserve contextual detail.
• Programmatic API: Exposes a RESTful API for programmatic queries and data retrieval.

Scientific Applications:

• Disease research: Supports analysis of protein-phenotype relationships relevant to rare and complex diseases, aiding identification of potential biomarkers and therapeutic targets.
• Biocuration: Provides a literature-derived source for validating and enriching databases with protein-phenotype associations.
• Literature mining: Enables locating and contextualizing articles that report protein-phenotype relationships to support hypothesis generation.
• Text-mining development: Serves as a dataset for training and evaluating predictive models that extract biological relationships from unstructured text.

Methodology:

Extracts protein-phenotype mentions from Medline and PubMed Central Open Access and analyzes co-occurrences within textual contexts (sentence and paragraph spans).