Co-expression analysis in close physical proximity using tumor data

There is increasing evidence that gene order within the eukaryotic genome is not random[1], and the genes that have a similar function have a tendency to express concurrently (co-expression). We analyzed expression data from six eukaryotic species (yeast, fruit fly, worm, rat, mouse and human) using publicly available datasets[2]. In the six species, there was a consistent tendency that neighboring genes on chromosomes were likely to be co-expressed. These results suggested the existence of some mechanisms for co-expressed genes in close physical proximity. In normal tissue, it is considered that co-expression relates higher-order function such as MAPK signaling pathway. In the tumor tissue, it is considered that biological system is different from normal tissue. If co-expression in close physical proximity relates higher-order functions such as cell division or proliferation, behavior of co-expression may change in the tumor tissue. The purpose of this work is to investigate whether the co-expression of genes in close physical proximity is also observed in the tumor tissue and what kinds of genes are co-expressed if it is observed. Using the huge clinically annotated tumor data, we calculated the relationship between co-expression rate (CoER) of gene pairs and the chromosomal distance (CD). We also investigated the Gene ontology categories enriched in the genes that highly correlated with multiple genes on the same chromosome in close physical proximity.