mQTL analysis
Log10 transformed data together with the tomato linkage map containing 865 SNP markers were used for performing QTL analysis using Rqtl v3.3.1 (Arends et al. 2010; Broman et al. 2003). We performed QTL analysis using the approach as previously described by Joosen et al., (2013) for Arabidopsis and Kazmi et al., (2017) for tomato, with small modifications. A model (Y = G + E + G×E + Ɛ) was used for the whole RIL population to identify the effect of genotype (G), maternal environment (E) and their interaction (G×E). Furthermore, in order to identify the metabolic variation explained by the genetic component we carried out the QTL analysis for known metabolites in each sub-population with simple interval mapping (SIM) using MapQTL® 6.0 (Van Ooijen 2004). In both QTL analysis (Rqtl and MapQTL) 1000 permutation tests were applied to our data to estimate the LOD threshold at a significance level of 0.05.