Review of Integrative Analysis Challenges in Systems Biology

Abstract

In the pharmaceutical industry, systems biology is making a paradigm shift in drug discovery from target-focused to a systems approach. It studies the molecular level changes in a biological system as an integrated and interacting network of genes, proteins, and metabolites instead of focusing on individual components. Integrative analysis of molecular level data becomes very important in unraveling the complex relationship and interactions in the biological system. One way to conduct integrative analysis is to identify correlations among genes, proteins, and metabolites. There are various analytic challenges: data preprocessing from various omics platforms; the number of variables much larger than the number of subjects; wide differences in the number of variables per platform leading to imbalance in the platform contributions to the analysis; the issues of multicollinearity and multiple testing; and difficulty in proper model validation to avoid selection bias. This article is a comprehensive review of statistical methods that address analysis challenges in integrative data.