LPInsider

LPInsider extracts long non-coding RNA–protein interactions (LPIs) from biomedical literature using text-mining features and machine learning to enable identification and analysis of LPI relationships.

Key Features:

• Automated Extraction: Extracts LPIs from biomedical texts using multiple text features including semantic word vectors, syntactic structure vectors, distance vectors, and part-of-speech vectors.
• Machine Learning Integration: Employs logistic regression to analyze extracted features and predict interactions, with performance evaluated across various feature combinations and models.
• High-Quality Corpus: Utilizes a manually filtered and reliable LPI corpus as a reference dataset for extraction and validation.

Scientific Applications:

• LncRNA Functional Characterization: Facilitates identification of lncRNA–protein associations to support functional analysis of lncRNAs.
• Disease Mechanism and Therapeutic Target Discovery: Provides extracted LPI data to support investigation of disease mechanisms and potential therapeutic targets.
• Literature Curation and Dataset Generation: Enables large-scale curation of LPIs from literature and generation of validated interaction datasets for downstream analyses.

Methodology:

Extraction from biomedical literature using semantic word vectors, syntactic structure vectors, distance vectors, and part-of-speech vectors; logistic regression for interaction prediction; use of a manually filtered LPI corpus; performance evaluated across feature combinations and models.