Table 1.
Breast cancer risk association results from logistic regression and mixture models of population training samples
| N | Logistic regression model | Mixture model | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Risk group | Variantsa | Cases | Controls | ORb | 95% CIc | P-value | Missense OR (95% CI)d | αe | 95% CIf |
| ATM | Log-likelihood = − 48,624.97 | Log-likelihood = − 48,624.64 | |||||||
| Non-carriers | – | 33,351 | 37,001 | 1 | – | – | 0 | – | |
| Carriers | 2.16 (1.78–2.63)h | ||||||||
| Variant outside FAT and PIK domains | 714 | 1259 | 1443 | 0.98 | (0.91–1.06) | 0.67 | 0.0041 | (0.001–0.02) | |
| Variant inside FAT or PIK domain and CADD score quintiles 1–4 g | 171 | 317 | 333 | 1.10 | (0.94–1.29) | 0.24 | 0.055 | (0.03–0.12) | |
| Variant inside FAT or PIK domain and CADD score quintile 5 g | 103 | 239 | 162 | 1.64 | (1.33–2.02) | 3.1 × 10−6 | 0.54 | (0.41–0.68) | |
| BRCA1 | Log-likelihood = − 48,652.14 | Log-likelihood = − 48,652.29 | |||||||
| Non-carriers | – | 34,191 | 37,996 | 1 | – | – | 0 | – | |
| Carriers | 10.61 (7.92–14.21)h | ||||||||
| Variant outside RING and BRCT domains | 479 | 811 | 856 | 1.01 | (0.92–1.12) | 0.79 | 0.0015 | (9.4 × 10−5–0.025) | |
| Variant inside RING or BRCT domain and low Helix score | 79 | 120 | 103 | 1.18 | (0.90–1.55) | 0.23 | 1.0 × 10−11 | NA | |
| Variant inside RING or BRCT domain and high Helix score | 23 | 63 | 16 | 4.94 | (2.83–8.61) | 1.9 × 10−8 | 0.48 | (0.19–0.78) | |
| BRCA2 | Log-likelihood = − 48,641.97 | Log-likelihood = − 48,638.78 | |||||||
| Non-carriers | – | 33,006 | 36,517 | 1 | – | – | 0 | – | |
| Carriers | 5.87 (4.75–7.24)h | ||||||||
| Variant with low Helix score | 1160 | 2062 | 2323 | 0.98 | (0.92–1.04) | 0.47 | 5.1 × 10−5 | (2.4 × 10−9–0.52) | |
| Variant with high Helix score | 62 | 114 | 94 | 1.28 | (0.96–1.70) | 0.087 | 0.11 | (0.04–0.25) | |
| CHEK2 | Log-likelihood = − 48,728.96 | Log-likelihood = − 48,728.70 | |||||||
| Non-carriers | – | 34,582 | 38,480 | 1 | – | – | 0 | – | |
| Carriers | 1.75 (1.47–2.08)i | ||||||||
| Variant with low Helix score | 157 | 403 | 363 | 1.26 | (1.08–1.46) | 0.0025 | 0.33 | (0.25–0.43) | |
| Variant with high Helix score | 121 | 265 | 177 | 1.73 | (1.42–2.11) | 4.7 × 10−8 | 0.95 | (0.86–0.98) | |
| PALB2 | Log-likelihood = − 48,728.67 | Log-likelihood = − 48,729.17 | |||||||
| Non-carriers | – | 34,622 | 38,291 | 1 | – | – | 0 | – | |
| Carriers | 424 | 618 | 713 | 0.95 | (0.85–1.06) | 0.34 | 4.87 (3.50–6.77)h | 1.1 × 10−4 | (1.6 × 10−9–0.88) |
a Number of unique missense substitutions in population dataset
b Logistic regression odds ratio estimate for missense variant carriers
c 95% confidence interval for logistic regression OR estimate for missense variant carriers
d Mixture model odds ratio and 95% confidence interval for missense variant carriers
e Alpha: estimated proportion of risk associated missense variants
f 95% confidence interval for alpha
g CADD quintiles 1–4 includes all CADD score values ≤ 3.736542; CADD quintile 5 includes all CADD score values > 3.736542
h Missense variant odds ratio constrained to equal odds ratio for protein truncating variants
i Missense variant odds ratio unconstrained