1061

We used the R/qtl software for the linkage map construction (Broman 2010). Only the markers with a chi-square less than 9.2 were kept for the analysis, the others ones were declared as distorted markers. The analysis included 101 samples and 73 markers: Among the markers analyzed, we had 9 putative genes markers, 43 ESSR, 18 GSSR and 3 BSSR. Data file is presented in Appendix 6.

We used the function est.rf (data) to estimate the recombination fraction between each pair of markers and we calculated with this same function the LOD score for a test of recombination fraction of 0.5. The results obtained were used to create the linkage groups with the function formLinkageGroups( ). We chose as arguments a maximum recombination fraction of 0.35 and a minimum LOD score of 2. Within the linkage groups, the markers were ordered with the function order.Markers ( ) and we finally obtained the linkage map presented in Appendix 7. Distance between markers were calculated in centi Morgan

We also constructed a plot of the linkage map with R (using Kosambi function) but the image obtained was too complicated to be presented here. In order to present a clearer plot of the map, we also construct linkage map with the software JoinMap (using Kosambi function) but we keep the results of the linkage map constructed by R to do the QTL analysis.

The map presented on this page is the R QTL analysis. For the JoinMap+Kosambi map see SS1061.

We also used R/qtl software to do the QTL analysis. Composite interval mapping with Haley-Knott regression (Haley and Knott, 1992) is usually considered as the more powerful method for QTL analysis but our phenotypic data did not fit a normal distribution. As we could not use this method in our case, we used a non parametric interval mapping. We used the function scanone ( ) and we chose one thousand as permutation number. QTL analysis was performed for seven phenotypic traits: lycopene, phytoene, xanthophyll, α-carotene, β-carotene, ζ-carotene and γ-carotene.