The Integrative Correlation Coefficient: a Measure of Cross-study

Leslie M. Cope, Johns Hopkins University
Liz Garrett-Mayer
Edward Gabrielson, Departments Oncology and Pathology, Johns Hopkins Medical Institute
Giovanni Parmigiani, The Sydney Kimmel Comprehensive Cancer Center, Johns Hopkins University & Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health

Abstract

Multi-study analysis adds value to microarray experiments. However, because of significant technical differences between microarray platforms, and because of differences in study design, it can be difficult to combine data. We have developed a statistical measure of reproducibility that can be applied to individual genes, measured in two different studies. This statistic, which we call the Integrative Correlation Coefficient or Correlation of Correlations, borrows strength across many genes to estimate the strength of the relationship between expression values in the two studies.