AbstractThe three‐dimensional structure of DNA depends subtly on its sequence, and this property is used by the proteins that regulate gene expression to locate their target sequences. Despite the large body of experimental data that has been accumulated on the relationship between sequence and DNA structure, we still do not fully understand the molecular basis for these properties, nor can we predict a three‐dimensional structure from a given sequence. We have been using computer modelling to tackle these problems. Some of our results and the implications for understanding the biological role of key sequences are discussed here.