However, identifying the traits that share a common association signal is not a trivial statistical task. Visual comparison of overlaps of association signals with an expression dataset is a step in this direction (using for example Sanger tool Genevar http://www.sanger.ac.uk/resources/software/genevar/), but the abundance of eQTLs in the human genome and across different tissues makes an accidental overlap between these signals very likely [2]. Therefore visual comparison is not enough to make inferences about causality and formal statistical tests must be used to address this question.
