Isolation of mutants in populations of yeast and other microorganisms, has been a valuable tool in experimental genetics for decades. The main disadvantages of classical mutant hunts are the inability to isolate mutants in non-selectable phenotypes and the difficulty of obtaining multiple mutations affecting a single phenotype, i.e. the isolation of mutants in polygenic traits. Most traits of organisms, however, are polygenic and non-selectable. This includes many, if not most, commercially-important traits of industrial yeast strains. The advent of powerful technologies for polygenic linkage analysis of complex traits now allows efficient identification of multiple mutations responsible for a complex trait among many thousands of irrelevant mutations. In this study, we have optimized a methodology for introducing hundreds of mutations into a single haploid S288c strain using multiple rounds of EMS mutagenesis, while maintaining genetic proficiency. Two mutants with about 900 mutations were screened for multiple non-selectable phenotypes. One mutant showed strongly reduced ethyl acetate production in semi-anaerobic fermentations, while in the other increased isoamyl acetate production was observed. Since there are still one or more unknown enzymes responsible for ethyl acetate production, we have mapped the quantitative loci (QTLs) underlying this trait using pooled-segregant whole-genome sequence analysis. Further dissection of the QTLs identified induced non-synonymous single nucleotide polymorphisms in both CEM1 and PMA1 as being causative for low ethyl acetate production. The CEM1 gene encodes a beta-ketoacyl synthase that is involved in mitochondrial fatty acid synthesis. PMA1 is an essential gene that encodes an H+-ATPase and the cytosolic pH is primarily regulated by Pma1. We have also identified the TPS1 allele of S288c, present in the background of the mutant strain, to be causative in a QTL with a strong link to low ethyl acetate production. TPS1 encodes trehalose-6-P synthase that is involved in the synthesis of trehalose, a well-known storage carbohydrate and a stress protectant. Trehalose-6-phosphate has also been implicated in the regulation of the influx of sugars into glycolysis. Our results demonstrate that mutant strains can be obtained in a straightforward way that have a multitude of mutations randomly spread all over the genome. A collection of such saturated mutants could be used for phenotypic screens, in order to identify the genetic basis of a wide range of monogenic and polygenic phenotypes for which no selectable system can be devised. A major advantage of this method is that only a limited number of mutant strains has to be screened in order to identify with great probability at least one strain that is affected in the trait of interest. Moreover, the genes underlying the defect in the trait of interest can then be identified in a relatively straightforward manner. This method also allows identification of causative alleles of traits, which are actually absent in the multiple mutant strain because of one or more other mutations. Because of the fact that the different mutations segregate in the descendants after crossing with a wild type strain, they become visible in the segregants and can effectively be detected. The number of mutations that can be accumulated in a single yeast strain is therefore limited by the ability of the multiple mutants to cope with these mutations rather than by the possibility of observing the trait of interest in the mutant strain. In this way our method eliminates the most important limitations of the classical mutagenesis and selection methods. It is able to generate mutants that are deficient in many polygenic and non-selectable phenotypes. Moreover, the underlying causative alleles of these phenotypes can be identified in a relatively easy and straightforward manner.