From Big Data to Precision Cures

IMAGE: Transition landscape of in gene expression networks; identifying health (left most) and terminal AD (right most); intermediate states (center) include MCI (middle), early (top) & late (bottom) onset AD

IMAGE: Transition landscape of in gene expression networks; identifying health (left most) and terminal AD (right most); intermediate states (center) include MCI (middle), early (top) & late (bottom) onset AD

chang_pic.png

From Big Data to Precision Cures

 
 

Neurodegenerative diseases show a striking phenotypic diversity driven by a complex array of factors including genetics, aging, gender, environment and disease stage.

To capture this complexity, we use predictive computational systems biology network models that hybridize novel causality inference (bottom-up) algorithms with conventional top-down Bayesian networks.

Through these models, we are able to integrate pan-omics (genetic, genomics, proteomics, metabolomics), 

phenotypic, clinical and longitudinal data – allowing us to discover causal relationships hidden in the diseases multi-factorial etiology.

When applied to data from Alzheimer’s patients stratified by genetic risk factors, gender, and disease stage, our approach identified sex-specific metabolic markers of disease progression. In males these metabolic networks emerged as dysregulation of branch-chain amino acids while females showed decreases in a network of phosphatidylcholines.

Moving from the patient level to the single cell level, we were able to use our models to integrate multi-scale omics data from specific cell types in central nerve systems. Insights derived from these analyses are allowing us to identify and validate key driver therapeutic targets in cell-type specific way across the spectrum of age-associated neurodegenerative diseases.