Extraction of transcript diversity from scientific literature

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• Transcript diversity, driven by alternative splicing, enhances biological complexity.
• Information on alternative splicing mechanisms and functions is fragmented across scientific literature.
• A centralized knowledge base is needed to interpret high-throughput data and understand gene expression.

Purpose of the Study:

• To develop and apply a text-mining method for extracting alternative splicing information from MEDLINE.
• To create a comprehensive database (LSAT) of genes with alternative transcripts.
• To facilitate a quantitative understanding of tissue-specific gene expression mechanisms.

Main Methods:

• A composite text-mining approach was used to process the entire MEDLINE database.
• Information extracted includes tissue specificity, isoform number, causative mechanisms, functional implications, and detection methods.
• The LSAT database was semi-automatically generated from the mined data.

Main Results:

• Identified 959 instances of tissue-specific splicing.
• Alternative splicing is suggested as the primary mechanism for transcript diversity in the nervous system.
• Provided new annotations for 1,860 genes and assigned the MeSH term 'alternative splicing' to 1,536 abstracts.

Conclusions:

• The LSAT database offers a valuable resource for studying transcript diversity and alternative splicing.
• The findings highlight the significant role of alternative splicing in generating functional complexity, particularly in the nervous system.
• The study demonstrates the utility of text mining for curating biological knowledge and improving data interpretation.