Yeast may be smaller and less complicated than other eukaryotes, but don't count on them for having simpler genetics. This is well exemplified in a recent report in Nature by Steinmetz et al., who examined a quantitative trait that yields a high-temperature growth (Htg) phenotype in selected Saccharomyces cerevisiae isolates. The unexpected complexity of their results serves as a warning for mappers of quantitative trait loci (QTL) in other organisms.

In order to map the Htg trait, Steinmetz et al. crossed two strains of S. cerevisiae: one strain with the Htg trait and one strain without it. They then used microarray analysis to compare the entire genomes of, and to map recombination points in, Htg+ and Htg− progeny of this cross, allowing them to pinpoint a locus on chromosome XIV that appeared to be linked to the Htg trait. However, neither the extensive sequencing of this locus nor the comparison of gene expression levels in this region provided a clear answer as to what gene — or genes — underlie this mapped QTL.

So to overcome this problem, Steinmetz et al. used a novel strategy called reciprocal hemizygosity analysis, in which new diploid strains were created that contained a haploid deletion in one or the other parental chromosome at a single gene of interest. In this way, each allele of a gene could be examined separately for its individual contribution to the phenotype. For the chromosome XIV QTL, alleles at two genes, MKT1 and RHO2 , from the Htg+ parent contributed to the Htg phenotype in the progeny. But, surprisingly, one allele at a third gene, END3 , from the Htg− parent also contributed to the Htg phenotype — a finding that helps to explain why the initial cross between the two strains produced a hybrid with heterosis (or 'hybrid vigour').

Such interesting results from a 'simple' organism might have unwelcome implications for attempts to map QTL in more complicated organisms — such as in agricultural crops, mice and humans — in which most QTL mapping is now taking place. The authors suggest that current techniques are likely to miss QTL if their effects are too small or if they are produced in part by unexpected combinations of alleles between strains, as with the Htg phenotype.