Skip to main content

Table 4 GMDR analysis on the best SNP–SNP and gene- smoking status interaction models

From: Association of the interaction between interleukin-1β gene polymorphism and smoking status with the diabetic nephropathy risk in a Chinese Han population

Locus no.

Best combination

Cross-validation consistency

Training balanced accuracy

Testing balanced accuracy

p-values *

SNP- SNP interactions *

    

2

rs16944, rs3917356

7/10

0.6523

0.4785

0.32

3

rs16944, rs3917356, rs1143634

8/10

0.6353

0.4874

0.62

4

rs16944, rs3917356, rs1143634, rs1143627

6/10

0.5664

0.4956

0.88

Gene- smoking interactions **

    

2

rs16944, smoking

10/10

0.7688

0.5721

0.011

3

rs16944, rs3917356, smoking

8/10

0.7129

0.4956

0.26

4

rs16944, rs3917356, rs1143634, smoking

5/10

0.5268

0.4785

0.76

5

rs16944, rs3917356, rs1143634, rs1143627, smoking

6/10

0.5664

0.4852

0.81

  1. *Adjusted for gender, age, TC, LDL-C, SBP, DBP, BMI and smoking
  2. **Adjusted for gender, age, TC, LDL-C, SBP, DBP and BMI