MineICA Framework for Transcriptomic Data Analysis

This technical report presents MineICA, a framework designed for the storage and analysis of independent component analysis (ICA) applied to transcriptomic data. It facilitates the integration of additional molecular, clinical, and pathological data associated with samples and genes. The framework introduces a new class, IcaSet, which extends the eSet class from the Biobase package, enabling the storage of both inputs and outputs from ICA. MineICA enhances the biological interpretation of components by examining their associations with various variables and biological processes. It also allows for the comparison of components across different datasets using correlation-based graphs. The report details the functionalities of MineICA, including storage of ICA results, analysis parameters, variable associations, feature annotations, and visualization capabilities. Furthermore, a case study demonstrates the application of MineICA using microarray-based gene expression data from breast cancer tumors, illustrating the process of loading libraries, creating IcaSet objects, and running ICA algorithms.