Development of informatics tools for complex gene systems: Killer-Cell immunoglobulin-like receptor genes

: Killer-cell immunoglobulin-like receptors (KIRs) are a newly described family of polymorphic and highly homologous genes that have been difficult to classify and characterize. Before comprehensive analyzes of the genes are completed, researchers must struggle with completing the task of classifying and characterizing what is currently known. A collection and alignment of all KIR sequences found in GenBank was created to facilitate oligonucleotide reagent development and to provide an overall picture of this complex gene system. Two methods, a direct measurement of homology and phylogenetic analysis, were used to categorize sequences previously not specifically identified as belonging to a particular locus. The two methods agreed for 64.2% of sequences. A further 22.6% of uncategorized sequences were specified by only one method, although the assignments were consistent. Some sequences (11.3%) could not be assigned to a locus by either method. For one sequence, the two methods disagreed as to the locus assignment (1.9%). The failure to categorize a sequence was usually related to the short length of the sequence and the similarity among KIR loci. The tools developed in this study have been valuable in the analyses of KIR sequences and can be used for any complex gene system.