Initial data integration and harmonization
The first step was to define where to start. Where were they missing information? Where does information need to be cleaned up? This was defined through a series of workshops to allow an effective prioritization on how to approach such a large journey of data transformation. ONTOFORCE began operating within the customer’s enterprise ecosystem, considering the connection to both upstream and downstream systems and integrating with necessary security measures. A wide range of sources were integrated, including internal relational databases like Oracle, public databases and API connections from external data providers such as Clarivate and Informa, as well as graph databases and different types of graph plugins that allowed the customer to automate a variety of processes. This was all incorporated into a platform that was fully branded according to the customers own internal IT architecture.
Identification of medical products (IDMP)
The first use case in the regulatory domain was an IDMP case (identification of medical products) which is a new strategy driven by the European regulator EMA. As guidelines for IDMP evolved, ONTOFORCE combined the various data domains together in the graph database according to the five relevant ISO standards. Then the DISQOVER non-GXP layer was added on top, providing connections and mappings between all the different data points. This is not only a richer layer, but it is easier and faster to bring it together because it is not subject to the same controls and regulations as the GXP system.
Data democratization was enabled through a unique way of extracting the data from the underlying operational layer, adding in the data model, and applying semantics and ontologies in a way that allows the non-data experts to query and retrieve relevant and complete information. This means that the data can be navigated effectively across the entire pharmaceutical value chain in a much more automated and coherent way. In gaining this visibility of the data quality across internal systems and data, as well as visibility of licensed data, it became clear where further investment and effort was needed for data cleanup.
In silico target identification
A further beneficial outcome of this data integration was a project based around in silico target identification. Information on genes, proteins, and diseases was aligned and analyzed in downstream algorithms. The benefit of DISQOVER here was to harmonize and provide a clear overview of the data, while offering control and flexibility in data modeling that was specifically tailored to their needs, offering much more visibility and transparency into how the data were harmonized than would a standard service solution. That target identification is combined with a competitive landscaping to provide insights into what can be done from a biological and chemical perspective and also where it makes economic and scientific sense to focus, based on the company’s specific expertise.
Case Study
