Figure 1

Approach proposed in this paper. Schematic representation of the analysis of genetical genomics experiment, the blue circles represent the markers tested and the yellow circles, each of the transcript levels analysed, an arrow signifies that the effect has been included in the model for the transcript. The left cartoon (a) represents the current strategy for eQTL searching: it consists of including the most significant marker in the model when testing each transcript independently. Several arrows pointing to a transcript means that the transcript is affected by several QTL, while many arrows starting in a single marker represents an eQTL hotspot (H). The right cartoon (b) presents the strategy proposed here, which suggests that external expression levels can be included as covariates in the model for the expression level studied (the arrows that start and end at the cDNA circles). Including cDNAs in the model can dramatically affect the final eQTL map, some positions may be shifted, some previous eQTL may disappear or some new appear. The bottom line of this approach is that all markers and all expression levels are potential regressors to be considered. The optimum model could be chosen using some of the available criteria, like AIC, BIC, DIC or AUC among others.