BioLT™ Linguistics Tool

BioLT leverages both a lexical and natural language processing approach to linguistics and text mining. Though BioLT can be configured to text mine any corpus of public or proprietary text-based repositories, it is currently focused on MEDLINE abstracts and being used as the automated component producing the NCI Cancer Gene Database.

By combining biological and medical term dictionaries with powerful free-text querying capabilities, BioLT allows the user to quickly, accurately, and efficiently access information that is often difficult to find in MEDLINE abstracts. The BioLT system can be configured to support clinical studies in complex disease research to facilitate diagnostic progress, chemical-compound development, and target-gene identification and validation. Doctors can easily access the latest diagnostic and treatment information to support patient care.

• Uses automatically generated and highly curated biological and medical term dictionaries.
• Can be customized according to topic-specific dictionaries and different text bodies.
• Includes 260,000 disease terms, 130,000 human gene names and name variants, 82,000 compound terms, 61,000 pathway terms, and more.
• Extracts relationships from a query expression, including synonyms of terms, using a powerful query language (including Boolean operators, dictionary entries, synonyms, wildcards and topics).
• Can automatically create vocabularies for any domain.
• Includes several ranking methods (chi-square, cosine, etc.).
• Finds the specific sentences in abstracts that demonstrate the relationship between compounds shown to effect complex diseases such as Alzheimer's.
• Supports cancer, cardiovascular, neurological, and infectious disease research, and more.
• Employs an easy-to-use Web interface for the user
• Is a UNIX client/server-based system.