Dataset Information

Humans and machines in biomedical knowledge curation: hypertrophic cardiomyopathy molecular mechanisms' representation.

ABSTRACT:

Background

Biomedical knowledge is dispersed in scientific literature and is growing constantly. Curation is the extraction of knowledge from unstructured data into a computable form and could be done manually or automatically. Hypertrophic cardiomyopathy (HCM) is the most common inherited cardiac disease, with genotype-phenotype associations still incompletely understood. We compared human- and machine-curated HCM molecular mechanisms' models and examined the performance of different machine approaches for that task.

Results

We created six models representing HCM molecular mechanisms using different approaches and made them publicly available, analyzed them as networks, and tried to explain the models' differences by the analysis of factors that affect the quality of machine-curated models (query constraints and reading systems' performance). A result of this work is also the Interactive HCM map, the only publicly available knowledge resource dedicated to HCM. Sizes and topological parameters of the networks differed notably, and a low consensus was found in terms of centrality measures between networks. Consensus about the most important nodes was achieved only with respect to one element (calcium). Models with a reduced level of noise were generated and cooperatively working elements were detected. REACH and TRIPS reading systems showed much higher accuracy than Sparser, but at the cost of extraction performance. TRIPS proved to be the best single reading system for text segments about HCM, in terms of the compromise between accuracy and extraction performance.

Conclusions

Different approaches in curation can produce models of the same disease with diverse characteristics, and they give rise to utterly different conclusions in subsequent analysis. The final purpose of the model should direct the choice of curation techniques. Manual curation represents the gold standard for information extraction in biomedical research and is most suitable when only high-quality elements for models are required. Automated curation provides more substance, but high level of noise is expected. Different curation strategies can reduce the level of human input needed. Biomedical knowledge would benefit overwhelmingly, especially as to its rapid growth, if computers were to be able to assist in analysis on a larger scale.

SUBMITTER: Glavaski M

PROVIDER: S-EPMC8487578 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:The vast collection of biomedical literature and its continued expansion has presented a number of challenges to researchers who require structured findings to stay abreast of and analyze molecular mechanisms relevant to their domain of interest. By structuring literature content into topic-specific machine-readable databases, the aggregate data from multiple articles can be used to infer trends that can be compared and contrasted with similar findings from topic-independent resources. Our study presents a generalized procedure for semi-automatically creating a custom topic-specific molecular interaction database through the use of text mining to assist manual curation. We apply the procedure to capture molecular events that underlie 'pain', a complex phenomenon with a large societal burden and unmet medical need. We describe how existing text mining solutions are used to build a pain-specific corpus, extract molecular events from it, add context to the extracted events and assess their relevance. The pain-specific corpus contains 765 692 documents from Medline and PubMed Central, from which we extracted 356 499 unique normalized molecular events, with 261 438 single protein events and 93 271 molecular interactions supplied by BioContext. Event chains are annotated with negation, speculation, anatomy, Gene Ontology terms, mutations, pain and disease relevance, which collectively provide detailed insight into how that event chain is associated with pain. The extracted relations are visualized in a wiki platform (wiki-pain.org) that enables efficient manual curation and exploration of the molecular mechanisms that underlie pain. Curation of 1500 grouped event chains ranked by pain relevance revealed 613 accurately extracted unique molecular interactions that in the future can be used to study the underlying mechanisms involved in pain. Our approach demonstrates that combining existing text mining tools with domain-specific terms and wiki-based visualization can facilitate rapid curation of molecular interactions to create a custom database. Database URL: •••