Biomarkers discovery in Parkinsonism from longitudinal and multi-modal big data sets of genetic, clinical, biological and imaging data

Speaker: Faraz Faghri,

Graduate Research Assistant,

Advisor: Roy Campbell

Abstract:

Parkinsonism syndrome has been used to describe neurologic disorders characterized by the presence of at least three of the following: tremor, rigidity, gait disturbance, and bradykinesia. Even though Parkinson’s disease (PD) is the most common cause of Parkinsonism, differential diagnosis of PD from other Parkinsonian disorders, such as Essential Tremor (ET), Multiple System Atrophy (MSA), Corticobasal Degeneration (CBD) and Progressive Supranuclear Palsy (PSP), can be clinically challenging, particularly during the early disease stages. The Parkinson’s disease itself is highly variable, with age of onset, rate of progression, and type and severity of symptoms different across the 5 million worldwide living with the disease.

We propose algorithms identifying models and biomarkers for prognosis and sub-typing which aids in subject selection for clinical studies and design of trials toward novel therapies. Integration of complex multidimensional datasets including clinical, biological, genetic, and imaging has proven to be a promising approach to help establish such discoveries. Our methods have identified various subtypes of the disease and their statistically significant patterns from clinical and biological data. Currently, we are extending the methods to incorporate TeraBytes of multi-modal image and genetic data. Our evaluated algorithms also empower a cloud-enabled web service to transform the healthcare community with assisted data-driven decision making tools.