1. Undesirable absorption, distribution, metabolism, excretion properties are the cause of many drug development failures and this has led to the need to identify such problems earlier in the development process. This work highlights computational (in silico) approaches used to identify characteristics influencing the metabolism of uridine diphosphate (UDP)-glucuronosyltransferase (UGT) substrates. Uridine diphosphate-glucuronosyltransferase facilitates conjugation between glucuronic acid and a nucleophilic site within a substrate and is one of the major drug-metabolizing enzymes. 2. An understanding of the relevant structural and chemical characteristics of the ligand and the enzyme active site will lead to greater utilization of metabolically relevant structural information in drug design. However, an X-ray crystal structure of UGT is not yet available, little has been reported about important structurally or catalytically relevant amino acids and only recently has the reported substrate profile of UGT isoforms reached an interpretable level. 3. A database of all the known substrates and non-substrates for each human UGT isoform was assembled and a range of modelling approaches assessed. Currently, pharmacophore models developed using Catalyst (Accelrys, San Diego, CA, USA) indicate that substrates of the UGT1A family share two key hydrophobic regions 3 and 6-7 A from the site of glucuronidation in a well-defined spatial geometry. Furthermore, two-dimensional quantitative structure-activity relationship models show significant reliance on substrate lipophilicity and a range of other descriptors that are known to capture information relevant to ligand-protein interactions. 4. In conclusion, substrate-based modelling of UGT appears both useful and feasible, with significant potential for determining aspects of chemical structure associated with metabolism and to quantify the nature of the relationship for UGT substrates. The development of a novel, user-defined 'glucuronidation feature' for alignment was crucial to the development of pharmacophore-based UGT models.