BER extracts both 15 key types of biological entities with subcategories leading to more than 40 distinctions and 17 relationships binding them, as well as rich contextual information. Extracted entities include Anatomical terms, Cells and Tissues (distinguishing between Body Fluids, Cell Types, Cell Tissues, Cell Anatomy and Cell Populations), Proteins and Genes (distinguishing between Enzymes, Antibodies, and Protein Genes), Disorders, Treatments, Species (including strain information), Genomic and Proteinic Information (distinguishing between Genomic and Proteinic Regions and microRNAs) as well as mutational information (including SNPs). Relationships include Activation, Inhibition, Regulation, Gene expression, Mutation, Diagnosis, and Therapy.
BER is powered by a deep syntactical analysis layer that provides an intimate and highly structured understanding of complex biological mechanisms and pathways. Biological entities extracted by BER are normalized by default according to MeSH, thus providing a MeSH ID number.
Customization and Extension
Thanks to its modular structure, BER offers unique customization capabilities, enabling the bespoke assembly of its internal Skill Units into bespoke extractors focused on novel areas of interest. On project basis BER may also be customized to normalize extracted entities according to alternative reference nomenclatures on demand.
BER is is well suited to process scientific literature and enable :
- the enrichment of biomedical literature with deep, domain-specific metadata
- the extraction of structured biological information to feed knowledge bases on key research areas
- the analysis of such to derive scientific insights pertaining to biological pathways, proteins/genes and their associations with diseases, drugs and their effects
For earlier versions of this Skill Cartridge®, please contact TEMIS.