Fig. 1.

Sample size flow chart.

Legend: Description of analytic sample with available genetic, smoking and lung cancer data for (A) Women's Health Initiative (WHI) Single Nucleotide Polymorphism Health Association Resource (SHARe) and (B) Multicenter Lung Cancer Case–Control Study. There were 8421 women in the downloaded dbGaP data for WHI SHARe (dbGaP accession #4723: Gene by environment interactions for lung cancer in cohort of African-American women in Women's Health Initiative), and 1265 were excluded for missing salient variables. The final WHI SHARe cohort included 7156 women with 59 cases of lung cancer and 29 cases of lung cancer mortality. There were 1,358 lung cancer cases and 1,289 controls in the multicenter case–control study of lung cancer. We excluded 50 lung cancer cases and 419 controls because of missing data, leaving 1,308 lung cancer cases and 1,241 controls as the final study population. When modeling interactions, never-smokers were excluded, leaving 1078 cases and 822 controls.