LION LBD: a literature-based discovery system for cancer biology

Abstract

Motivation The overwhelming size and rapid growth of the biomedical literature make it impossible for scientists to read all studies related to their work, potentially leading to missed connections and wasted time and resources. Literature-based discovery (LBD) aims to alleviate these issues by identifying implicit links between disjoint parts of the literature. While LBD has been studied in depth since its introduction three decades ago, there has been limited work making use of recent advances in biomedical text processing methods in LBD. Results We present LION LBD, a literature-based discovery system that enables researchers to navigate published information and supports hypothesis generation and testing. The system is built with a particular focus on the molecular biology of cancer using state-of-the-art machine learning and natural language processing methods, including named entity recognition and grounding to domain ontologies covering a wide range of entity types and a novel approach to detecting references to the hallmarks of cancer in text. LION LBD implements a broad selection of co-occurrence based metrics for analyzing the strength of entity associations, and its design allows real-time search to discover indirect associations between entities in a database of tens of millions of publications while preserving the ability of users to explore each mention in its original context in the literature. Evaluations of the system demonstrate its ability to identify undiscovered links and rank relevant concepts highly among potential connections. Availability and implementation The LION LBD system is available via a web-based user interface and a programmable API, and all components of the system are made available under open licenses from the project home page http://lbd.lionproject.net. Supplementary information Supplementary data are available at Bioinformatics online.

SKiM - A generalized literature-based discovery system for uncovering novel biomedical knowledge from PubMed

“…The form of LBD we implement here is the “open discovery A➔B➔C model,” which takes only a concept of interest (the “A” term) and retrieves a list of associated concepts (“C” terms) through some intermediate concepts (“B” terms). The application of LBD to the scientific community is limited due to the lack of influential methodology 9,10 . Existing LBD systems fall into one of the following categories: co-occurrence models, semantic models, graph-based models, distributional models, and hybrid models 7,11 .…”

“…Existing LBD systems fall into one of the following categories: co-occurrence models, semantic models, graph-based models, distributional models, and hybrid models 7,11 . While many LBD systems have been built 7,8,10,12 , only BITOLA 8 and LION LBD 10 have a functioning web interface. BITOLA 8 is restricted to use the semantic types from the Unified Medical Language System (UMLS) Metathesaurus 13 as B and C terms.…”

“…Drug repurposing with BITOLA is incomplete, as the available controlled vocabularies do not contain a comprehensive list of drugs and do not include many drugs and synonyms from DrugBank 14 and PharmGKB 15,16 . LION LBD is specific to cancer biology, and restricted to use only chemicals, diseases, mutations, genes, cancer hallmarks, and species as B and C terms 10 . The system is not flexible to run on specific types of concepts as B (e.g.…”

“…SKiM builds upon the features of KinderMiner 19 , our recently developed text mining system to filter statistically significant co-occurring concept pairs from PubMed abstracts. Many existing LBD systems 10,20,21 and popular text mining systems such as PolySearch2 2 use a co-occurrence approach for extracting information from PubMed. KinderMiner gave interesting and promising results on identifying transcription factors for cell reprogramming, retrieved potential drugs for lowering blood glucose 19 , performed better than STRING 22 and PolySearch2 2 on retrieving mitochondrial protein-protein interactions (PPI) 23 , identified lab tests that can be repurposed for other diagnostic goals 24 , and identified phenotypes associated with FMR1 premutation 25 .…”

Inhibition of the Dead Box RNA Helicase 3 prevents HIV-1 Tat and cocaine-induced neurotoxicity by targeting microglia activation

“…can rapidly reveal hidden connections in the literature network (20,21). Machine learning approaches accelerate drug repositioning to discover new applications for existing compounds (22).…”