Table 3 Comparison of logic regression and benchmark methods to detect 3-SNP haplotypes in the discovery cohort, N = 1,937.
From: Genome-wide haplotype association analysis of primary biliary cholangitis risk in Japanese
Chr | Method A (proposed): Logic regression | Method B (benchmark): 3-SNP sliding windows | Comparison | |||||
|---|---|---|---|---|---|---|---|---|
# Gene windows with replicated 3-SNP haplotype (Method A) | Gene window with best p-value | Best p-value | # Tests with P < 3.2 × 10−8 (Bonferroni) | # Gene windows with at least one haplotype with P < 3.2 × 10−8 | Gene window with best p-value | Best p-value | # Gene window matches between methods A and B | |
2 | 0 | NA | NA | 1 | 1 | LRP1B | 1.1 × 10−9 | 0 |
3 | 0 | NA | NA | 5 | 3 | NEK10 | 3.6 × 10−18 | 0 |
6 | 143 | NOTCH4 | 6.1 × 10−27 | 1352 | 173 | TAAR2 | 2.2 × 10−30 | 123 |
7 | 9 | GLCCI1 | 8.7 × 10−6 | 6 | 3 | HGF | 2.1 × 10−15 | 0 |
8 | 0 | NA | NA | 4 | 4 | CYP11B1 | 1.1 × 10−8 | 0 |
9 | 12 | DEC1 | 2.5 × 10−17 | 74 | 16 | DEC1 | 3.2 × 10−13 | 12 |
18 | 0 | NA | NA | 10 | 5 | MTCL1 | 1.6 × 10−10 | 0 |
