Figure 4.

eQTLs controlling the Glucosinolate (GS) network gene expression. The vertical lines separate the genome into chromosomes I-V (labeled at the top) with cM progressing left to right along the x axis. The length of each chromosome in cM is indicated at the end of each chromosome. A. Network eQTLs detected (LOD profiles) for the glucosinolate network gene expression estimated as the average of the log₂ expression across all glucosinolate genes (mean^log2 = red profile), average the log₂ expression across only the glucosinolate genes minus those with a large cis effect eQTL (mean^-cis = blue profile) and the average normalized expression across all glucosinolate genes (mean^z = black profile). The threshold for declaring a QTL significant at α = 0.05 is LOD > 2.8. B. Allele effect trace for the glucosinolate network gene expression estimates shown in A. The y axis is in either log₂ or z units, depending upon the profile. The effect shown is in terms of the Bay-0 allele. A positive value indicates that the Bay-0 allele has a positive effect on the trait. A negative value indicates that the Bay-0 allele has a negative effect while the Sha allele has a positive effect on the trait. C. eQTLs controlling the log₂ expression of the individual genes in the glucosinolate network. The color schematic at the bottom shows the LOD score scale with the directionality of effect indicated by the color. Only significant QTLs are shown. An eQTL for which the Bay-0 allele has a positive effect is shown by differing shades of yellow and red, and darker intensity indicates a greater LOD score at that genetic position. An eQTL where the Bay-0 allele has a negative effect is shown by differing shades of blue, again with darker intensity indicating a greater LOD score. Gene names in red indicate those with a cis-eQTL that controls > 50% of the phenotypic variation for that gene.