Sofia – Assertional Metadata Creation
Sofia enables the creation of networks of Assertional Metadata (AMD), a form of Intelligent Metadata that summarises the assertions made within unstructured or poorly formatted text and providing a semantically consistent method of navigation. Sofia liberates intelligence from structured and unstructured data, regardless of its original format.
The AMD Intelligence Networks made possible by the Sofia platform are a new type of knowledge representation, a format that enables the integration of knowledge from a wide range of diverse sources in a semantically consistent way so that information can be easily navigated, analyzed, shared and exploited. Creating Intelligence Networks at the scale and accuracy required to support pharmaceutical business decisions needs a robust methodology, powerful automation and stringent quality control. BioWisdom developed the Sofia environment in partnership with some of the worlds leading pharmaceutical companies where the technology is being exploited to address real-world scientific and commercial issues.
SofiaEditor is a powerful tool for knowledge professionals which provides editorial control over the creation of Intelligence Networks and incorporates all of the processes involved in acquiring, normalizing, and validating intelligence for inclusion in an Intelligence Network. sofiaEditor has been designed to work at scale, with huge structured databases, flat file and semi-structured data sources. Information from unstructured data sources such as Medline, regulatory documents etc. can be extracted using sofiaEditor’s intelligent text-mining features enabled by tailored Key Concept Metadata (KMD). Assertions that cannot be mined automatically can be extracted using sofiaEditor’s semi-automated document markup features.
Intelligence Networks are extensible and can be reused in multiple applications over time. Multiple returns on investment in Intelligence Networks are enabled through the sofiaServer API’s which provides programmatic access to information in the Intelligence Network.
SofiaBrowser provides the main interface through which Intelligence Networks are viewed, queried and exploited. SofiaBrowser allows information to be navigated and queried using text strings or chemical structures. SofiaBrowser enables large-scale querying across Intelligence Networks by name, relationship type, property or category. SofiaBrowser’s visualizations include clustered graph displays and interactive categorical selections. SofiaBrowser also provides links to popular analytical tools such as Excel, Spotfire, SAS and a range of chemoinformatics systems. SofiaBrowser also provides pathway searching tools to allow connected sets of concepts to be found in the network.
BioWisdom’s KCM is used to power used Sofia’s ability to automatically read millions of documents and to create metadata links to specific assertions of interest made within them. Sofia Vocabulary Sets are also used to organize, navigate and export assertions for use in other applications.
Tags: AMD, Assertional Metadata, Intelligence Networks, Sofia