In this study, Saccharomyces cerevisiae was used to develop a method for predicting phenotype from genomic sequence. The phenotypic impact of sequence polymorphisms found in 19 strains was predicted computationally based on functional information and data from reverse genetic screens. Strains were ranked with respect to the likelihood of being affected for a particular phenotype; these predictions were then tested by assessing the growth rate and efficiency of 15 strains grown under 20 conditions. Predictions were highly accurate, especially for gene sets that were functionally connected.