Table 1.
False positive report probability (FPRP) values for four results on associations between 16 variants in genes involved in the repair of double-stranded DNA breaks and breast cancer based on data in Kuschel et al. (25)
| Gene/SNP | Odds ratio (95% CI)* | Statistical power under recessive model† | Reported P value‡ | Prior probability§ |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| .25 | .1 | .01 | .001 | .0001 | .00001 | |||||||||
| XRCC3 C→T at nt 18067 | 1.32 (1.08 to 1.60) | 1.00 | .015 | .042 | .12 | .59 | .94 | .993 | .9993 | |||||
| XRCC3 A→G at nt 17893 | 0.82 (0.72 to 0.94)‖ | .9895 | .0075 | .022 | .064 | .43 | .88 | .987 | .9987 | |||||
| LIG4 T→C at nt 1977 | 0.65 (0.42 to 0.98) | .87 | .088 | .23 | .48 | .91 | .990 | .9990 | 1.00 | |||||
| XRCC2 G→A at nt 31479 | 2.60 (1.00 to 6.73) | .17 | .071 | .56 | .79 | .98 | .998 | .9998 | 1.00 | |||||
| XRCC3 haplotype | 1.00¶ | .000016# | .000049 | .00015 | .0016 | .016 | .14 | .62 | ||||||
Odds ratios, except as noted, were calculated for the homozygotes with rare genetic variants versus the referents for homozygotes with common genetic variants, as reported in Table 2 of (25). CI = confidence interval. SNP = single-nucleotide polymorphism.
Statistical power, except as noted, is the power to detect an odds ratio of 1.5 for the homozygotes with the rare genetic variant (or, 0.67 = 1/1.5 for protective effect) and 1 for the heterozygotes and for the homozygote with the common variant, with an α level equal to the reported P value.
P values were calculated using the omnibus chi-square test, with two degrees of freedom, as reported in Table 2 of (25). The FPRP values are based on these P values.
The most likely range of prior probabilities are in bold type for each gene/SNP or haplotype. The prior probability is for an effect of the gene/SNP in the direction of the observed odds ratio.
Odds ratios were calculated for the heterozygotes with the genetic variants versus the referent for the homozygotes with the common genetic variants, as reported in Table 2 of (25).
Statistical power to reject the null hypothesis using the omnibus chi-square test, when the odds ratio is 1.5 for the second most frequent haplotype and 1 for the other haplotypes, with the most common haplotype as referent.
P value was calculated using the omnibus chi-square test, with seven degrees of freedom.