INTRODUCTION
IMGT/GENE-DB, part of IMGT, the international ImMunoGeneTics information system®, http://imgt.cines.fr (1–4) is the comprehensive IMGT genome database, which has been developed to classify the immunoglobulin (IG) and the T cell receptor (TR) genes from vertebrate species, and to standardize and manage the complex IG and TR gene data knowledge (5) (http://www.bioinfo.de/isb/2003/04/0004/). The molecular genetics of the IG and TR genes is so complex and unique in the genome of vertebrates (6,7) that a specific gene database was required to manage all their characteristics. Indeed, the synthesis of IG and TR chains involves multigene families from four different gene types: variable (V), diversity (D), joining (J) and constant (C), each one with unique characteristics. These genes are organized in hundreds of cassettes, as in fish, or in large clusters from several hundred kilobases to one (or more) megabase(s), as in mouse and human (6,7). IG and TR genes that belong to same subgroup may be highly similar in their coding sequence, but at the same time, highly polymorphic (e.g. 13 allelic forms have been sequenced for the human IGHV2-70 gene) (6), with alleles displaying different functionalities. The presence of many pseudogenes in the loci, and the frequency of the polymorphisms by gene insertion and deletion in these multigene families, add an additional level of complexity (6,7). Although most human IG and TR genes were sequenced and characterized independently from and before the completion of the Human Genome Project, the classification and the characterization of the IG and TR genes remain a big challenge in the analysis of the genome. Indeed, the annotations of the IG and TR loci, which represent for instance, in human, ∼6 Mb on chromosomes 2, 7, 14 and 22, are not available through classical genome software, owing to the unique IG and TR gene structure (6,7). At the level of gene expression analysis (e.g. cDNAs), data are even more difficult to interpret as the mechanisms involved in the IG and TR synthesis include DNA rearrangements with large DNA deletion of several hundred kilobases, and recombinations, nucleotide deletions and insertions at the rearranged junctions and, for IG, somatic hypermutations. Such somatic mechanisms create an extraordinary diversity of 1012 different IG and TR per individual (6,7). Thus, most IG and TR expressed sequences, available in IMGT/LIGM-DB (8) (http://www3.oup.co.uk/nar/database/summary/504), the IMGT sequence database, and in IMGT/3Dstructure-DB, the IMGT 3D structure database (9) show significant nucleotide and amino acid differences, respectively, by comparison with the germline (not rearranged) sequences. IMGT/GENE-DB has been implemented to provide an easy and common access to standardized and expertly annotated IG and TR gene and allele data and knowledge. The first task of IMGT was to define a reference sequence for each individual gene and allele (6,7), based on the IMGT ‘gene’ and ‘allele’ concepts. IMGT/GENE-DB has been developed using Java and cgi programs and has been available on the Web since January 2003. IMGT/GENE-DB, which currently contains human and mouse IG and TR genes, is the international reference for the IG and TR gene nomenclature.