Chagas cardiomyopathy prediction using statistical models.
Project Status
Complete
Project Deliverables
To analyze and leverage the data that Dr.Seidman’s group has on hand to better understand progression to Chagas cardiomyopathy
Project Description
CCB worked with Seidman Lab to help process, explore, and analyze the data that Dr. Seidman’s lab has collected. This involved data cleaning as appropriate and re-examining the epitope quantification pipeline and batch effect correction steps. We tested a suite of standard machine learning techniques (using random forests as a starting point) to develop a predictive classifier that uses the epitope data and other available covariates to distinguish between the Chagas cardiomyopathy and indeterminate patient categories
Outcomes
We reviewed and updated the Seidman group’s QC pipeline on a PhIP-Seq assay focused on cardiomyopathy patients. The newly corrected data was used as input to a variety of machine learning models aimed at predicting disease status.