Two accessions, representing the species Lycopersicon esculentum (cultivated tomato) and Lycopersicon pennellii (a wild relative), were evaluated for 11 quantitative traits and found to be significantly different for 10 of the traits. Transgressive segregation was observed for eight of the traits in a large interspecific F2 population. When restriction fragment length polymorphism markers were used as probes for the quantitative trait loci (QTL) underlying the traits, 74 significant QTL (LOD > 2) were detected. Thirty-six percent of those QTL had alleles with effects opposite to those predicted by the parental phenotypes. These QTL were directly related to the appearance of transgressive individuals in the F2 for those traits which showed transgressive segregation. However, the same types of QTL (with allelic effects opposite to those predicted by the parents) were also observed for traits that did not display transgressive segregation in the F2. One such trait was dry weight accumulation. When two overdominant QTL (detected in the F2) for this trait were backcrossed into the L. esculentum genetic background, transgressive individuals were recovered and their occurrence was associated with the two QTL demonstrating the potential for transgressive segregation for all characters and implicating overdominance as a second cause of transgressive segregation. Epistasis was not implicated in transgressive segregation in either the F2 or backcross generations. Results from this research not only reveal the basis of wide-cross transgressive segregation, but demonstrate that molecular markers can be used to identify QTL (from wild species) responsible for transgressive phenotypes and to selectively transfer them into crop species. This strategy might be used to improve many traits of economic importance including those for which wild species appear phenotypically inferior to their cultivated counterparts.
