Genome annotation is the application of useful biological descriptions to sequence data. Different levels of time and effort can be invested to produce correspondingly different depths of annotation depending on what methods are employed. Researchers using genome data should, therefore, understand how annotations are generated to assess their validity correctly and to determine what level of inferences can be made accordingly. Thorough annotation requires a large range of procedures, most of which involve manual reviews of all available evidence. First, gene structures often are computed algorithmically and edited based on in-depth analyses of the underlying sequence data. Second, functional predictions draw on data from various sources. Finally, the use of structured and controlled descriptions, such as those provided by gene ontology, can be used so that final descriptions are not only consistent and unambiguous, but capable of being used in further downstream analyses such as cross-species comparisons.