Salt tolerance in Lycopersicon species. V. Does genetic variability at quantitative trait loci affect their analysis?

Salt tolerance was studied comparatively in three families derived from crosses between Lycopersicon esculentum Mill. and two related wild species [two accessions of Lycopersicon pimpinellifolium (Jusl.) Mill. and one accession of Lycopersicon chesmannii f. minor (Hook. f.) Mull.] by means of QTL analysis of fruit yield and earliness under conditions of salinity. From six polymorphic genomic regions involved in salt tolerance, three contained segregant salt-tolerant QTLs for the three families; two were found only in both families derived from L. pimpinellifolium; and one, involved in fruit number, was detected only in one of the L. pimpinellifolium families. Some differences regarding the effects of the wild alleles at orthologous QTLs were found. These effects were always negative in the L. chesmannii family. Comparing both L. pimpinellifolium families, the "wild" alleles at two out of nine common QTLs for fruit number and weight had effects with opposite directions, and the mode of gene action was clearly different at five of them. QTL analysis of earliness revealed the largest genotypic differences among families. Most drastic differences were found for the epistatic interactions in which all genomic regions containing QTLs were involved. These interactions between unlinked genes increased the range of variation of means, mainly upwards, as compared with genotypes at individual QTLs. Only one (affecting fruit weight) out of 27 interactions was detected in both L. pimpinellifolium families. Heterotic effects found for salt tolerance in one of the families can be explained by the presence of overdominant (or pseudo-overdominant) and dominant gene effects at QTLs controlling final fruit yield under conditions of salinity. Allelic variation at salt-tolerant QTLs exists, changing the additive and, mainly, the non-additive components of the genotypic value. Consequently, it may negatively affect the general applicability (or efficiency) of marker-assisted selection to improve salt tolerance in other segregant populations where QTLs were not studied. The use of more informative co-dominant markers, like microsatellites, might overcome these problems.