Phenotypic QTL Enrichment Analysis

To investigate whether genomic regions containing more phenotypic QTLs per region than expected by chance are associated with the network eQTLs, we conducted a sliding window analysis and compared the results (Figure 1 – Summation Approach, using Phenotypic QTL instead of eQTL). A diverse range of 62 biochemical, morphological and plant/biotic interactions traits were included (Additional file 3) [29 (link),30 (link),34 (link),48 (link)-52 (link)]. The 62 traits identified 281 phenotypic QTLs based on the 38 microsatellite marker map for 411 Bay-0 × Sha RILs [34 (link)], resulting in an average of 4.5 QTLs per trait. Because these data were measured on all 411 RILs, we used the highest-resolution map available for this RIL collection. The empirical threshold for a significance level of 0.05 for the frequency of traits with a phenotypic QTL per 5 cM sliding window was estimated as described above.