[Application of multifactor dimensionality reduction on the interactions between gene-gene, gene-environment and the risk sporadic colorectal cancer in Chinese population].

Objective: To identify the association between risk of sporadic colorectal cancer and the common single nucleotide polymorphisms (SNPs) in DNA repairs genes, gene to gene interactions among them and their gene to environment interactions with common environmental factors.

Methods: In this population-based case-control study, 206 primary colorectal cancer cases and 845 cancer-free healthy controls were enrolled. Genotyping was carried out using polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) technique, with the status of subjects case or controls unknown. Multifactor dimensionality reduction (MDR) and logistic analysis were both used for association analysis.

Results: As compared to the younger age group (> or = 42, < 61 years), the risk of colorectal cancer in older age group (> or = 61 years) increased significantly (OR = 2.04, 95% CI: 1.49-2.80). Similar result was observed in the family cancer history (OR = 1.51, 95% CI: 1.05-2.17). However, no significant association between any single DNA repair gene SNP and colorectal cancer risk was discovered. Results from MDR analysis only showed a significant interaction among the four following factors: age, alcohol drinking, XRCC1 Arg194Trp and OGG1 Ser326Cys (the cross-validation consistency = 10/10, the average testing accuracy = 0.616, P = 0.011). Using a logistic regression model, the "high-risk" individuals had a significantly elevated risk of colorectal cancer compared to those "low- risk" individuals classified by the above MDR model (OR = 2.72, 95% CI: 1.66-4.47).

Conclusion: The impact of polymorphisms in DNA repair genes on the risk of sporadic colorectal cancer exhibited a low-penetrance characteristics while the intricate interactions existing among them and with environmental factors.