Article Text

Download PDFPDF
Massive false-positive gene–gene interactions by Rothman’s additive model
  1. Kwangwoo Kim
  1. Department of Biology, Kyung Hee University, Seoul 02447, South Korea
  1. Correspondence to Professor Kwangwoo Kim, Department of Biology, Kyung Hee University, Seoul 02447, South Korea; kkim{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Rothman’s additive interaction model based on additivity of the relative risks of two disease-risk factors has frequently been used to investigate gene–gene interactions for various traits, especially for rheumatic diseases compared with other diseases (online supplementary text). For example, a recent paper by Diaz-Gallo et al 1 reported significant additive interactions between the HLA-DRB1 shared-epitope allele and numerous non-HLA rheumatoid arthritis (RA) risk variants (eg, PTPN22 variants) in anti-citrullinated protein antibody–positive (ACPA-positive) RA.

Supplemental material


This study investigated the validity of the additive interaction model in assessing gene–gene interactions. Genome-wide association study (GWAS) datasets including 1000 cases and 1000 controls were simulated 100 times based on an actual GWAS dataset from 2849 WTCCC2 controls (1958_British_Birth) to have European genomic features and a single variant with OR≈5, 3, 2, 1.5, 1.2 or 1 in a dominant genetic model (used in a typical additive interaction analysis). The simulated dataset is expected to have no true disease-risk effects of variants (except for an OR-simulated variant) and no true biological interaction effects between any variants on risk of a fake disease. Using the simulated datasets, all possible interactions between an OR-simulated variant and other variants in different chromosomes …

View Full Text


  • Handling editor Josef S Smolen

  • Contributors KK designed the study, performed analyses and wrote the manuscript.

  • Funding This research was supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT and Future Planning (2017R1E1A1A01076388) and the Korea Healthcare Technology R&D Project funded by the Ministry for Health and Welfare (HI15C3182) in Republic of Korea.

  • Disclaimer The Consortium and/or Individual Investigators bear no responsibility for the further analysis or interpretation of these data, over and above that published by the Consortium.

  • Competing interests None declared.

  • Patient consent Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.