Global Similarity and Local Variance in Human Gene Coexpression Networks

For the study presented here, we performed a comparative analysis of whole-genome gene expression variation in 210 unrelated HapMap individuals to assess the extent of expression divergence between 4 human populations and to explore the connection between the variation of gene expression and function. We used the GEO series GSE6536 to compare changes in expression of 47,294 human transcripts between four human populations. Gene expression patterns were resolved into gene coexpression networks and the topological properties of these networks were compared. The interrogation of coexpression networks allows for the use of a well-developed set of analytical and conceptual tools and provides an opportunity for the simultaneous comparison of variation at different levels of systemic organization, i.e., global vs. local network properties. The results of this comparison indicate that human co-expression networks are indistinguishable in terms of their global properties but show divergence at the local level.