PosMed

PosMed prioritizes candidate genes and other biomedical entities by ranking statistical associations between user-defined phenotypic keywords and genes, metabolites, diseases, and drugs using a Semantic Web Association Study engine that integrates MEDLINE, OMIM, pathway, co-expression, molecular interaction, and ontology data.

Key Features:

• Semantic Web Association Study engine: Ranks genes, metabolites, diseases, and drugs by statistical significance of associations between phenotypic keywords and biological entities.
• Integrated data sources: Integrates MEDLINE, OMIM, pathways, co-expression data, molecular interaction data, and ontology terms from twenty public databases and four original datasets.
• Direct and inferential associations: Establishes direct and inferential links between phenotypic keywords and entities via an extensive network of biological databases.
• Text-mining with human curation: Applies text-mining rules developed through human curation to generate extensive gene–document linkages.
• Positional cloning and linkage prioritization: Supports in silico positional cloning and prioritization of candidate genes within chromosomal intervals derived from linkage analysis.
• Exome sequencing interpretation: Enables functional interpretation and prioritization of genetic variants identified by exome sequencing of human disease samples.
• Cross-species coverage: Performs association searches across species including human, mouse, rat, and Arabidopsis thaliana.

Scientific Applications:

• Gene prioritization for positional cloning: Prioritizing candidate genes within linkage or chromosomal intervals for disease gene discovery.
• Functional interpretation of exome variants: Interpreting and prioritizing genetic variants from exome sequencing in human disease studies.
• Linking mouse bioresources to phenotypes: Linking mouse bioresources lacking phenotypic annotations to relevant phenotypes via inferential gene–document associations.
• Cross-species association discovery: Performing cross-species association searches to leverage evidence from human, mouse, rat, and Arabidopsis thaliana.
• Drug and metabolite association ranking: Ranking drugs and metabolites by association to phenotypic keywords to support hypothesis generation.

Methodology:

Uses a Semantic Web Association Study engine that ranks associations by statistical significance using direct and inferential links derived from integrated resources (MEDLINE, OMIM, pathways, co-expression, molecular interactions, ontology terms), applies human-curated text-mining rules to build gene–document linkages, and integrates twenty public databases plus four original datasets for cross-species association searches.