iTextMine

iTextMine integrates multiple text-mining tools to extract, align, and standardize relations and entities from Medline abstracts and PMC open access full-length articles for large-scale biomedical knowledge extraction.

Key Features:

• Automated Workflow: Enables simultaneous execution of multiple text-mining tools on large-scale datasets.
• Parallel Processing with Dockerization: Uses parallel processing and dockerized tool instances to enable consistent execution across computing environments.
• Standardized JSON Output Format: Standardizes outputs from diverse tools into a uniform JSON format for aggregation and downstream analysis.
• Text Alignment Algorithm: Implements a novel text alignment algorithm to resolve discrepancies among tool outputs and map extracted information to source text.
• Integration of Multiple Relation Extraction Tools: Integrates four distinct relation extraction tools to process all Medline abstracts and PMC open access full-length articles.

Scientific Applications:

• Database curation: Automates extraction and integration of literature-derived relations to support biological database curation.
• Hypothesis generation: Identifies literature-based associations to facilitate generation of testable hypotheses.
• Gene–disease association analysis: Enables analysis of gene-disease associations demonstrated with genes PTEN and SATB1 in breast cancer.

Methodology:

Automated workflow for simultaneous tool execution; parallel processing via dockerized tool instances; standardization of outputs into JSON; application of a novel text alignment algorithm to reconcile tool outputs; integration of four relation extraction tools applied to all Medline abstracts and PMC open access full-length articles.