Associations with breast and ovarian cancer risk for SNPs observed at p-trend<0.05 in stage II of the experiment.

a<p>Hazard Ratio per allele (1 df) estimated from the retrospective likelihood analysis.</p>b<p>Hazard Ratio under the genotype specific models (2df) estimated from the retrospective likelihood analysis.</p>c<p>p-values were based on the score test.</p>d<p>HR per allele of 1.69 and p-trend of 1×10⁻⁴ for <i>BRCA2</i> mutation carriers in stage I of the study.</p>e<p>HR per allele of 1.43 and p-trend of 0.01 for <i>BRCA1</i> mutation carriers in stage I of the study.</p>f<p>HR per allele of 1.30 and p-trend of 0.03 for <i>BRCA1</i> mutation carriers in stage I of the study.</p>g<p>HR per allele of 0.64 and p-trend of 0.057 for <i>BRCA2</i> mutation carriers in stage I of the study.</p>h<p>HR per allele of 1.25 and p-trend of 0.04 for <i>BRCA1</i> mutation carriers in stage I of the study.</p>i<p>HR per allele of 1.25 and p-trend of 0.058 for <i>BRCA2</i> mutation carriers in stage I of the study.</p>j<p>rs3093926 did not yield results under the genotype specific model due to the low minor allele frequency.</p><p>Complete description of results from stage I are included in Supplementary <a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1004256#pgen.1004256.s002" target="_blank">Table S1</a>.</p><p>Highlighted in bold are those SNPs showing strongest associations with breast or ovarian cancer risk (p<0.01).