Haplotype | Frequency* (%) | Effect | P^{‡} | |||||
---|---|---|---|---|---|---|---|---|

Rs925013/rs266882 | Controls, N = 734 | Cases, N = 819 | OR^{†} (95% CI) | |||||

Global score statistic | 0.02 | |||||||

Additive^{§} | ||||||||

A/A | 52 | 49 | Reference | |||||

A/G | 30 | 28 | 1 (0.8-1.2) | 0.9 | ||||

G/A | 1 | 1 | 1.7 (0.7-3.8) | 0.2 | ||||

G/G | 18 | 22 | 1.3 (1.1-1.6) | 0.008 | ||||

Dominant^{∥} | ||||||||

A/A | 52 | 49 | Reference | |||||

A/G | 30 | 28 | 0.9 (0.7-1.1) | 0.4 | ||||

G/A | 1 | 1 | 1.5 (0.7-3.6) | 0.3 | ||||

G/G | 18 | 22 | 1.3 (1.1-1.7) | 0.009 | ||||

Recessive^{¶} | ||||||||

A/A | 52 | 49 | Reference | |||||

A/G | 30 | 28 | 1.1 (0.7-1.5) | 0.8 | ||||

G/A | 1 | 1 | NA | NA | ||||

G/G | 18 | 22 | 1.3 (0.8-2.2) | 0.2 |

↵* Estimated haplotype probabilities estimated from genotype data using the haplo.cc function of the HaploStats library in R.

↵† ORs obtained using recursively the estimated posterior probabilities of pairs of haplotypes per subjects as weights in the logistic model (haplo.cc function of the HaploStats library in R).

↵‡ The

*P*for the global score statistic corresponds to the test for overall association between haplotypes and prostate cancer. The other*P*s based on the score statistics (Schaid et al., 2002) correspond to the test for association between the specific haplotype and prostate cancer risk.↵§ In the additive model, it is assumed a linear effect of the number of copies of the haplotype (0, 1, and 2). Therefore, ORs estimate the risk for men carrying one copy of the haplotype relative to the risk for men homozygous for the reference haplotype (A/A). The ORs for men carrying two copies of the haplotype can be obtained by applying the following formula: exp[2 × log(OR

_{1})], where OR_{1}, the OR for men carrying one copy of the haplotype, is reported in the table.↵∥ In the dominant model, the ORs estimate the risk for men heterozygous or homozygous for the haplotype relative to the risk for men homozygous for the reference haplotype (A/A).

↵¶ In the recessive model, the ORs estimate the risk for men homozygous relative to the risk for men heterozygous (for the same haplotype) or homozygous for the reference haplotype (A/A).