METHODS Subjects Lymphocyte DNA was analysed from patients with familial pancreatic cancer enrolled in the National Familial Pancreatic Tumor Registry. 15 Patients with pancreatic cancer were selected if they had at least two or more first degree relatives with pancreatic cancer (mean (SD) age of 66.7 (12.3) years, males 50.3%). Variants were analysed in 110 additional patients with familial pancreatic cancer. To determine the carrier frequency of c. 2574C. G (p. Ser868Arg), we analysed three control populations: healthy spouses of patients with familial pancreatic cancer (115 samples from spouses with a mean (SD) age of 66.9 (11.3) years, males 43.1%), patients who had undergone cholecystectomy (65 samples matched in age with sporadic cases) for non-malignant disease at Johns Hopkins Hospital, and individuals undergoing routine screening colonoscopy (668 samples) at the Mayo Clinic. The mean age of the colonoscopy controls was similar to our pancreatic cancer population (mean (SD) age of 59.3 (12.3) years, males 52.9%). Lymphocyte DNA was obtained from spousal controls and colonoscopy controls, and formalin fixed paraffin embedded DNA from the resected gallbladders of hospital controls. None of the controls had a prior history of pancreatic cancer or colorectal cancer. We also analysed germline DNA from 124 individuals with apparently sporadic pancreatic cancer (those without a known family history of pancreatic cancer; mean (SD) age of 66.23 (10.3) years, males 48%). This study was approved by the Johns Hopkins Joint Committee for Clinical Investigation and the Mayo Clinic IRB.

FANCA analysis FANCA (Genbank# NM_000135. 1) was analysed using a combination of heteroduplex analysis and DNA sequencing. A detailed description of methods used including primer sequences can be accessed at http://www. pathology2. jhu. edu/pancreas/FANCA and at http://jmg. bmjjournals. com/supplemental/. The Fanconi Anemia Mutation Database