TERMite

TERMite is a high-performance named entity recognition (NER) and extraction engine developed by SciBite, designed to unlock vital information buried within complex collections of scientific text. It is built for scientists, researchers, and IT professionals working with large volumes of unstructured documents — including publications, patents, grants, electronic lab notebooks (ELNs), project management tools, and industry databases — who need to rapidly identify, tag, and organise critical scientific terms and relationships.

By pairing powerful processing technology with SciBite's hand-curated VOCabs — containing over 20 million synonyms across 80+ science topics — TERMite transforms unstructured content into rich, machine-readable data. The result is faster discovery of key facts, entities, and relationships across any document collection, without the need for pre-indexing or complex setup procedures.

Core Capabilities

• Ultra-fast processing: TERMite indexes at up to one million words per second, enabling users to start processing millions of documents within minutes of deployment.
• Massive scale: Capable of scanning billions of documents and handling large-scale document processing on distributed systems such as Hadoop.
• High accuracy: Precisely tags and links scientific terms within unstructured text using SciBite's VOCabs, reducing ambiguity and improving the reliability of extracted data.
• Up to 23 NER machine learning models: The latest version of TERMite incorporates up to 23 NER machine learning models to further enhance recognition accuracy.
• One-click CENtree integration: Enables direct editing of publicly available ontologies through seamless integration with SciBite's CENtree platform.

Key Use Cases and Benefits

• Data-mining at scale: Locate important information by identifying critical mentions and relationships across literature, patents, grants, and internal documents.
• Enhanced internal search: Augment existing search portals with more accurate entity recognition, improving team performance, productivity, and satisfaction.
• Content enrichment for diverse roles: Suitable for anyone producing or managing scientific textual content, including those supplying IT systems such as ELNs and industry databases, enabling improved search and navigation.
• Flexible workflow integration: Designed to fit seamlessly into existing analysis workflows, supporting both one-off analyses and routine automated processing as part of ongoing operations.

TERMite 6.6.3 — What's New

• New vocabularies introduced:
  • MONDO (Clinical Pack): A comprehensive disease ontology based on the Mondo Disease Ontology, harmonising disease definitions across susceptibility, injury, and disease branches.
  • EMTREE_PERSON (Emtree Pack): Focuses on named groups of persons including age and sex categories, derived from the 2024.03 version of Emtree.
  • UNIPROTMOUSE (GenPhen Pack): A new vocabulary of mouse proteins sourced from UniProt.
  • TAXPATH (Core Pack): Human pathogen taxonomy replacing the deprecated PATHOGEN vocabulary, with expanded coverage of over 23,000 entities.
  • TAXVERT (Core Pack): Human vertebrate taxonomy sourced from NCBI Taxon, version 2024-07-03.
• Key vocabulary updates:
  • Significant updates to HGNC, ChEMBL, Cellosaurus, MedDRA, and others, including new concepts, reorganised branches, and improved mappings to public ontologies.
  • Major expansion of the PKPD vocabulary, nearly doubling its size with new branches and terms.
  • Updated EMTREE branches covering procedures, healthcare, devices, organisms, and diseases, with hundreds of new concepts including SARS-CoV-2 variants and recent classifications.
• Bug fixes and security enhancements: Resolved issues related to batch processing and server validation errors, and addressed the latest CVEs to ensure platform security and stability.
• Deprecated vocabulary: The PATHOGEN vocabulary has been deprecated and replaced by the new TAXPATH vocabulary.
• Additional updates: Introduction of a new Meddevice pack for medical device vocabularies, starting with the MEDDEVICE vocabulary with plans for future expansion; the Help menu now includes a direct link to SciBite Academy for customer training.

TERMite is continuously updated to provide access to the latest product upgrades and expanded infrastructure support. It is supported by a formal Service Level Agreement framework and is complemented by SciBite resources including whitepapers on pharmacovigilance applications and RDF-based semantic data integration, making it a comprehensive solution for life sciences organisations seeking to maximise the value of their scientific text assets.