The identification of all protein targets of a small molecule drug, i.e. target deconvolution, provides the basis for understanding its beneficial or deleterious actions. Target deconvolution remains a major and often problematic task within the biotechnology and pharmaceutical industries; there is therefore a generally acknowledged need for robust alternative methods to complement existing techniques. This thesis presents the development and application of a novel yeast-based platform for small molecule target deconvolution. The platform couples a sensitive yeast three-hybrid screening system for the identification of potential target proteins to an independent validation step based on affinity chromatography. The development of a novel yeast three-hybrid system was primarily based on the use of SNAP-tag to covalently anchor small molecules of interest onto a bait protein. The sensitivity of the system was improved by using an engineered reporter yeast strain to increase SNAP-tag labelling efficiency in living yeast cells. The reporter strain was engineered by knocking out genes encoding broad-specificity transporters from the yeast genome in order to reduce small molecule efflux from yeast cells. Such engineering was also necessary for the development of a negative selection strategy to reduce false positive levels in cDNA libraries. Finally, the system was judiciously optimised using defined small molecule-protein interactions of known affinity. In the second part of this work, the newly developed yeast three-hybrid system was applied to the screening of cDNA libraries for small molecule target deconvolution. Thirty-five small molecules, including well-characterised approved pharmaceutical drugs, were screened against eight different human cDNA libraries, resulting in the identification of forty-one small molecule-protein interactions. These screening results comprised both previously known and unknown interactions. The identification of numerous known targets of approved drugs validates the performance of the yeast-based approach for target deconvolution. Additionally, several of the previously unknown interactions were further validated using affinity chromatography and activity assays. Amongst these validated interactions, this work has uncovered the first non-kinase target of the epidermal growth factor receptor (EGFR) inhibitor erlotinib, i.e. oxysterol-binding protein-related protein 7 (ORP7); it has also led to the identification of off-targets of the popular HMG-CoA reductase inhibitor atorvastatin. The most significant finding of this thesis is that the widely used anti-inflammatory drug sulfasalazine is a potent inhibitor of sepiapterin reductase, an enzyme involved in the biosynthesis of the cofactor tetrahydrobiopterin. Despite its long use and clinical importance in the treatment of inflammatory bowel diseases and rheumatoid arthritis, its mechanism of action is not well understood. Here, it is proposed that sulfasalazine and/or its metabolite(s) inhibit sepiapterin reductase in vivo, this leading to a decrease in tetrahydrobiopterin biosynthesis and to an associated anti-inflammatory effect. This hypothesis provides an attractive explanation for some of the drug's properties and opens possibilities for new and improved therapies. In summary, this work establishes a powerful approach for small molecule target deconvolution and provides new insights into the mechanism of action of clinically used drugs.