The transformation of unstructured text into contextualized data is facilitated through semantic enrichment and good ontology plays a critical role in the process. Contextualized data is required for various computational approaches such as Natural Language Processing, Text mining & analytics, machine learning. Even though ontologies form the central core of most semantic platforms, the process of maintaining them is a challenge for most organizations. Subject matter and ontology experts along with manual curation arerequired to maintain ontologies evolving from both public and proprietary sources.
Molecular Connections now provides a suite of enriched, linked biomedical ontologies accessible throughAPI technologies which are fast, flexible and easy to deploy. Making this a defacto component of anybiomedical data-led systems.
Molecular Connections has developed a repository of more than 80 open and commercial ontologies withinthe biomedical domain, which has resulted in an entity store of over 3.8 million biomedical concepts withhierarchical relationships spanning across multiple ontologies. In the process we have also considerably enriched the concepts with synonyms, public identifiers and other meta data by mining various public datasources.
How does it work?
Molecular Connections leverages machine learning complemented with manual curation to enrich and maintain updates to the ontologies in a controlled way. Molecular Connections proprietary ontology management solution “MCLEXICONTM” uses open standards to import and maintain ontologies which makes it easy to integrate the ontologies with other internal platforms. It also provides versioning of updates to the ontologies, with provisions to either accept or reject them. We also use machine learning approaches to disambiguate concepts across domains, by adding context to each concept. Ontologies can be accessed by Application Programing Interfaces (APIs) either using Molecular Connections proprietary text mining solutions (MCMINERTM) for Named entity recognition & extraction or by any other custom built third party solutions.
Speaking on the occasion Mr. Jignesh Bhate, CEO Molecular Connections mentioned, “through this service we aim to offer the biomedical science community a comprehensive suite of semantic solutions including Automated indexing, Natural Language Processing, Content Recommendation system, Content / Linked data management, Content Integration (for Competitive Intelligence) bringing efficiency, scale and discoverability in Content management.