Network graph, for Alzheimer’s disease concept, from Instem Scientific’s data integration platform. The lines radiating out from the centre show the relations from a set of Alzheimer’s disease concepts to interacting proteins, mRNAs and genes, pathology findings and other biomedical observations, and related tissues and anatomical structures. This harmonised set of data can be searched, navigated and analysed for significant interactions.
Biomarkers seem to be a major focus currently for the healthcare industry – with old ones continually being evaluated, and new ones being sought. There are two major types of biomarkers: biomarkers of exposure, for use in risk prediction, and biomarkers of disease, which are used in screening, diagnosis and monitoring of disease progression. We have recently published a paper jointly with King’s College London, that describes the discovery of two new potential protein biomarkers for Alzheimer’s disease (PLAUR and ChAt) using a combination of in silico prediction and in vitro validation.
For the in silico prediction, we used a powerful combination of life science vocabularies and our in-house data integration and data harmonisation technologies to pull together assertions linking Alzheimer’s disease pathologies, related proteins and anatomical structures. We could then analyse all the disease-related assertions in the intelligence network against findings for normal tissues, to pull out key proteins that could then be validated against a panel of samples from the European AddNeuroMed cohort.
This approach could be used for any disease state, toxicity or pathology – and shows the value of integrated harmonised access to legacy data – and of course, the benefits gained from good commercial-academic collaboration for the healthcare industry.
