Nature, 404, 398–402 (2000).

We reported the association of interleukin-1 gene cluster polymorphisms suspected of enhancing production of interleukin-1 β with an increased risk both of hypochlorhydria induced by Helicobacter pylori and of gastric cancer. We were subsequently alerted to an error in our genotyping by a report1 that the IL-1B-511T variant is in positive linkage disequilibrium with the IL-1B-31C allele, rather than with the IL-1B-31T allele as we had stated. We have traced this discrepancy to incorrect labelling of the sequence data from the wild-type and variant controls for the IL-1B-31 assays. Consequently, the designations IL-1B-31T and IL-1B-31C were reversed in the text and Tables 1–3. We have now verified the proper identifications by re-sequencing our controls. In addition, we have confirmed the more common (IL-1B-31T/IL-1B-511C) and less common (IL-1B-31C/IL-1B-511T) haplotypes by forward and reverse sequencing five homozygous samples (two wild-type and three variant) for both of these loci. Our data and analysis remain otherwise unchanged. The electrophoretic mobility shift assays were unaffected by this error and hence the greater DNA binding of the IL-1B-31T-bearing oligonucleotide represents the wild-type IL-1B promoter.