Analysis and improvements to Kabat and structurally correct numbering of antibody variable domains

In analysing protein sequence and structure, standardized numbering schemes allow comparison of features without explicit alignment. This has proved particularly valuable in the case of antibodies. The most widely used schemes (Kabat: sequence-based; Chothia: structure-based) differ only in the numbering of the complementarity determining regions (CDRs). We have analyzed the numbered annotations in the widely used Kabat database and found that approximately 10% of entries contain errors or inconsistencies. Further analysis of sequence alignments in the context of structure suggest that the sites of the insertions in some framework regions in the Kabat and Chothia schemes are incorrect. We therefore propose a corrected version of the Chothia scheme which is structurally correct throughout the CDRs and frameworks. To perform this analysis, we have developed, and made available, a tool for the automatic application of Kabat, Chothia and modified-Chothia numbering schemes and have carefully benchmarked the performance of this tool.