Current structural genomics projects are yielding structures for proteins whose functions are unknown. Accordingly, there is a pressing requirement for computational methods for function prediction. Here we present PHUNCTIONER, an automatic method for structure-based function prediction using automatically extracted functional sites (residues associated to functions). The method relates proteins with the same function through structural alignments and extracts 3D profiles of conserved residues. Functional features to train the method are extracted from the Gene Ontology (GO) database. The method extracts these features from the entire GO hierarchy and hence is applicable across the whole range of function specificity. 3D profiles associated with 121 GO annotations were extracted. We tested the power of the method both for the prediction of function and for the extraction of functional sites. The success of function prediction by our method was compared with the standard homology-based method. In the zone of low sequence similarity (Ϸ15%), our method assigns the correct GO annotation in 90% of the protein structures considered, Ϸ20% higher than inheritance of function from the closest homologue.functional residue ͉ function prediction ͉ structural genomics I ncreasingly, protein structures are being determined without a knowledge of the function of the molecule. Proteins with unassigned function began to accumulate in the Protein Data Bank (PDB) (1) 6 years ago, and their number is growing exponentially (see supporting information, which is published on the PNAS web site). Today, there are Ϸ500 proteins annotated as ''hypothetical'' in PDB (roughly 1 per 50 entries). This gap is expected to increase dramatically as a consequence of structural genomics projects in which high-throughput methods are applied to determine the conformations of numerous proteins in a genome-wide strategy (e.g., ref. 2). One major motivation of structural genomics projects is that the determination of the structure of a protein provides insight into its molecular function, which is a step toward understanding its cellular function. The current structure-function gap clearly shows that more powerful bioinformatics techniques for function prediction are urgently needed (3-5). Recently, several groups have developed algorithms to identify functionally important residues often employing sequence conservation and͞or structural information (see below). However, identification of function residues is distinct from actually assigning a function to the protein. Here we present an automated structure-based method for function prediction.The complexity of protein function makes the establishment of any functional classification problematic (6, 7). Today, an extensively used functional classification is derived from the Gene Ontology (GO) project (8). By means of GO, one can establish a functional hierarchy that progresses from general functions to more specific functions. As exemplified in GO, protein function ranges from the very general (e.g., en...