Barak Cohen and colleagues have developed a massively parallel synthetic promoter assay that they use to examine the effects of an array of transcription factor binding site variants (Genome Res. doi:10.1101/gr.157891.113, 6 August 2013). They used a modified cis-regulatory element sequencing approach that allows for chromosomal integration of large numbers of synthetic promoters in Saccharomyces cerevisiae. They constructed a barcoded reporter library that includes synthetic promoters with a combination of 12 transcription factor binding sites, for a total of 7,289 barcoded sequences representing 2,534 unique cis-regulatory elements. Three variants were included for each of four transcription factors—Mig1, Reb1, Rap1 and Gcr1. To examine a wide interaction space, the variants were selected to range from very low to very high predicted binding affinity. To analyze this data set, the authors used a thermodynamic framework that models both transcription factor binding to DNA and interactions between transcription factors. They found that their model replicated known transcription factor relationships and was able to explain 57% of the variance in expression in the promoter library. They also compared their thermodynamic model predictions to position weight matrix models for binding affinities based on analysis of data from chromatin immunoprecipitation and sequencing (ChIP-seq).