Computational Tool for Applications of Sparse Canonical Correlation Analysis on Biological Data

Abstract

Sparse canonical correlation analysis (sparse CCA) is a method for identifying sparse linear combinations of the two sets of variables that are highly correlated with each other, given that those two sets of measurements are available on the same set of observations. Recently, sparse CCA has become a popular method for analyzing genomic data, where the number of features is large compared to that of observations. Analyzing a set of data using sparse CCA requires multiple steps, including data cleaning, normalizing, and using the right programming packages.

To make sparse CCA accessible for all researchers regardless of their statistical background, a user-friendly computational tool should be created to assist them in walking through the analysis. After the tool is successfully implemented, a few sets of data will be used as case studies for testing efficiency of the sparse CCA computational tool. Eventually, the tool will be added to the computational website hosted by the Center for Translational Environmental Health Research, which currently hosts services for sequencing classification and differential expression analysis.