ONTOFORCE DISQOVER Natural Language Processing (NLP) (1)

DISQOVER NLP

DISQOVER enhanced with natural language processing (NLP)

Enhance DISQOVER with NLP

DISQOVER’s NLP technology comes with a powerful named entity recognition (NER) and an extraction engine which allows you to quickly find and analyze information about specific biological entities and their relationships (such as diseases, genes, proteins, and more), populations, interventions, and outcomes that are mentioned in large volumes of text data.

All with one single solution.

Key benefits of DISQOVER's NLP functionality

Pre-processed enhanced public data out-of-the-box
Process private data with pre-configured yet adjustable annotators
One-stop-shop: one single solution for structured AND unstructured data
Modular and flexible: deploy NLP for a single use case or for multiple use cases at once
Expert team of data scientists, machine learning engineers, and bioinformaticians available to support with implementation and user roll-out

Enhance DISQOVER with natural language processing packs

Available out-of-the-box for public data

Uncover insights and knowledge that were previously hidden in the text of articles and publications. Our NLP models have been trained to extract PICOS and biomedical entities and relationships from plain text. These are used to annotate and enrich DISQOVER’s public data, which is available via remote data subscription (RDS) packages that are ready to ingest without the need for any additional deployments.

The NLP enhancement for public data uses two models:

The PICOS model: retrieve population, intervention, comparison, outcomes, and study type data from the contents (such as descriptions and inclusion/exclusion criteria) of publications and clinical studies.
The Biomedical Entities model: distils biomedical data from the descriptions of clinical trials and literature.

Often, the publicly available clinical studies do not specifically annotate patient demographic, intervention, outcome, or biomarker information in a structured way, meaning important knowledge or studies may be missed. Through NLP, DISQOVER can reach this siloed knowledge, add it to the knowledge graph and make it available for search and exploration.

Apply on your private data

Find more hidden information inside your own, private data.

The NLP functionality for private data comes with the following three models:

The PICOS model: retrieve patient population, intervention, comparison, outcomes, and study type data from your internal documents as well as from the data in your own internal repositories such as clinical trial management systems.
The Biomedical Entities model: distils biomedical data from your internal document repositories.
The Informed Consent model: extracts relevant classification for primary and secondary usage of samples and data from consent forms to facilitate data reusability.

Available NLP enhanced packs for DISQOVER

NLP Enhanced Clinical Pack - Public Data

Data sources: Clinical studies, publications
Data types: Annotators for
1. Biomedical entities
2. PICOS criteria
Processing: Pre-processed
Implementation: Out-of-the-box, via RDS packages

NLP Enhanced Clinical Packs - Private Data

Data sources: Internal document repositories such as clinical trial management systems, informed consent forms
Data types: Annotators for
1. Biomedical entities
2. PICOS criteria
3. ICFs
Processing: Allows for the processing of private data only
Implementation: Customizations to your own data ingestion pipeline

Key features

Semantic identifiers for the most important concepts in public biomedical documents linked to standardized terminologies
Semantic search capabilities
Concept mapping: harmonized over a variety of vocabularies
Enhanced NLP for proprietary documents