One in seven people worldwide suffers from a brain disorder, e.g., epilepsy, Parkinson's, stroke, or dementia. Development of future treatments depends on improving our understanding of brain function and disease, and validating new treatments critically depends on identifying the underlying biomarkers associated with different conditions. Biomarker discovery requires volume, quality, richness, and diversity of data. This Direct-to-Phase II project extends Blackfynn's cloud data management platform for team science, in order to support interactive data curation and integration and to facilitate biomarker discovery. Our first technical aim develops tools to help select, curate, assess, and regularize datasets: we develop novel ?live? query capabilities to ensure users discover relevant data, develop mechanisms for using data's provenance to decide on trustworthiness, and build tools for mapping fields to common data elements. These capabilities address the critical, under-served problem of selecting the data to analyze. Our second technical aim develops techniques for incorporating algorithms to link and co-register across multi-modal data and metadata. Using ranking and machine learning, we can incorporate and combine state-of-the-art algorithms for finding data relationships, and we can link to remote data sources. These capabilities enable scientists to analyze richer datasets with multiple data modalities and properties ? thus enabling them to discover more complex correlations and biomarkers. In our third aim, Blackfynn's new technical capabilities will be applied to challenges faced by Blackfynn partners, including problems assessing trustworthiness of data annotations, conducting image analysis, modeling epileptic networks, and identifying biomarkers for neuro-oncology indications. As part of this validation we will also develop HIPAA-compliant mechanisms for working with protected and de-identified data together. Together, these three thrusts will ensure that development of the Blackfynn platform results in tools and technologies that meaningfully accelerate scientific understanding and discovery over rich and complex data, leading to improved treatments for neurologic disease.
This Direct-to-Phase II project extends the Blackfynn cloud data management platform to enable biomarker discovery for research and development of improved drugs, devices and clinical care for patients with neurologic disease: it develops tools for assembling, evaluating, and rating data, and linking it across modalities and to external systems. It also validates the techniques' effectiveness using real challenges faced by Blackfynn partners, in imaging, epilepsy, and brain tumor research.
