The identification of unknown compounds isolated during Effect-Directed Analysis (EDA) is often a hurdle on the way to the successful outcome of these studies. Ever-improving separation, analytical, and biological techniques allow the isolation of more compounds and effects; however, not all of the compounds contributing to sample toxicity are easily identified. The advancement of database search strategies and publishing of online databases has improved tentative identification of many compounds in recent years, but many chemicals and their transformation products are still not captured within such databases. Structure generation, where the analytical information is used to identify substructures present and absent, provides an alternative strategy to database searching. Where multiple structures matching an unknown spectrum are possible, candidate selection becomes critical to successful identification. The main steps in candidate identification and selection are discussed in this chapter, including examples of programs and strategies available. Improvements in the ability to share data between institutes and the selection criteria for candidate structures are needed to take advantage of recent analytical developments and further enhance structure elucidation in EDA studies.