Vapur

Vapur indexes protein–chemical pairs in CORD-19 abstracts to enable retrieval of studies linking proteins and chemicals relevant to COVID-19.

Key Features:

• Relation-Oriented Inverted Index: Constructs a semantic inverted index that groups and retrieves publications based on extracted relations between proteins and chemicals rather than keyword matches.
• BioNLP Pipeline Integration: Employs a BioNLP pipeline to perform named entity recognition and relation extraction on CORD-19 abstracts to populate the relation-oriented index.

Scientific Applications:

• Protein–chemical relationship discovery: Identifies literature reporting associations between specific proteins and chemicals within the COVID-19 corpus.
• Therapeutic target identification: Surfaces publications that link proteins and chemicals to support identification of candidate targets for COVID-19 research.
• Molecular interaction literature mining: Aggregates evidence of molecular interactions from CORD-19 abstracts for downstream bioinformatics analysis.

Methodology:

Named entity recognition and relation extraction are applied to CORD-19 abstracts via a BioNLP pipeline to build a semantic, relation-oriented inverted index.