REBASE is a comprehensive database containing information on restriction enzymes, DNA methyltransferases, and related proteins involved in restriction-modification (R-M) systems. It includes data on recognition and cleavage sites, isoschizomers, neoschizomers, commercial availability, methylation sensitivity, crystal and sequence data, and homing endonucleases. REBASE also contains information on putative R-M systems found in sequence databases and provides comprehensive descriptions of R-M content in fully sequenced genomes. The database can be browsed online (http://rebase.neb.com) and selected compilations can be downloaded via ftp (ftp.neb.com). Monthly updates can be requested via email. REBASE contains information on 3945 biochemically or genetically characterized restriction enzymes, with 299 distinct specificities among the 3834 Type II restriction enzymes. 641 restriction enzymes are commercially available, including 235 distinct specificities. The number of putative R-M systems is increasing rapidly, while the rate of biochemical characterization has decreased. However, the large number of sequenced examples of biochemically characterized restriction systems allows for the inference of putative recognition sequences. DNA methyltransferase genes are more reliable indicators of R-M systems. REBASE includes all likely candidates until it becomes clear that non-DNA methyltransferases have been included erroneously. REBASE is a valuable resource for the annotation of bacterial and archaeal genomes. It provides tools for analyzing sequences, including NEBcutter for finding restriction enzyme recognition sites and BLAST for searching against all sequences in REBASE. Specialty lists of sequence data are available for download. The coming year will see major additions to REBASE, including the inclusion of all metagenomic sequence data and a tool for users to perform their own analysis of newly sequenced genomes. REBASE is funded by the National Library of Medicine and New England Biolabs.REBASE is a comprehensive database containing information on restriction enzymes, DNA methyltransferases, and related proteins involved in restriction-modification (R-M) systems. It includes data on recognition and cleavage sites, isoschizomers, neoschizomers, commercial availability, methylation sensitivity, crystal and sequence data, and homing endonucleases. REBASE also contains information on putative R-M systems found in sequence databases and provides comprehensive descriptions of R-M content in fully sequenced genomes. The database can be browsed online (http://rebase.neb.com) and selected compilations can be downloaded via ftp (ftp.neb.com). Monthly updates can be requested via email. REBASE contains information on 3945 biochemically or genetically characterized restriction enzymes, with 299 distinct specificities among the 3834 Type II restriction enzymes. 641 restriction enzymes are commercially available, including 235 distinct specificities. The number of putative R-M systems is increasing rapidly, while the rate of biochemical characterization has decreased. However, the large number of sequenced examples of biochemically characterized restriction systems allows for the inference of putative recognition sequences. DNA methyltransferase genes are more reliable indicators of R-M systems. REBASE includes all likely candidates until it becomes clear that non-DNA methyltransferases have been included erroneously. REBASE is a valuable resource for the annotation of bacterial and archaeal genomes. It provides tools for analyzing sequences, including NEBcutter for finding restriction enzyme recognition sites and BLAST for searching against all sequences in REBASE. Specialty lists of sequence data are available for download. The coming year will see major additions to REBASE, including the inclusion of all metagenomic sequence data and a tool for users to perform their own analysis of newly sequenced genomes. REBASE is funded by the National Library of Medicine and New England Biolabs.