Documentation / Classification / Combining Rules
Combining Rules
Bayesian Point System
Helix Insight uses the Bayesian point-based classification framework (Tavtigian et al. 2018, 2020). Each evidence criterion contributes points based on its strength level. The total point sum determines the final classification.
Pathogenic Evidence Points
Benign Evidence Points
Classification Thresholds
| Classification | Point Range | Confidence Range |
|---|---|---|
| Pathogenic | >= 10 pts | 0.80-0.99 |
| Likely Pathogenic | 6 to 9 pts | 0.70-0.90 |
| VUS | 0 to 5 pts | 0.30-0.60 |
| Likely Benign | -1 to -5 pts | 0.70-0.90 |
| Benign | <= -6 pts | 0.80-0.99 |
ACMG 2015 Combining Rules (Reference)
The original 18 ACMG combining rules are a special case of the Bayesian point system -- every rule produces the same classification under both approaches. They are retained as a reference.
Pathogenic (8 rules)
P1: 1 Very Strong + >= 1 Strong
P2: 1 Very Strong + >= 2 Moderate
P3: 1 Very Strong + 1 Moderate + 1 Supporting
P4: 1 Very Strong + >= 2 Supporting
P5: >= 2 Strong
P6: 1 Strong + >= 3 Moderate
P7: 1 Strong + 2 Moderate + >= 2 Supporting
P8: 1 Strong + >= 4 Moderate
Likely Pathogenic (6 rules)
LP1: 1 Very Strong + 1 Moderate
LP2: 1 Strong + 1-2 Moderate
LP3: 1 Strong + >= 2 Supporting
LP4: >= 3 Moderate
LP5: 2 Moderate + >= 2 Supporting
LP6: 1 Moderate + >= 4 Supporting
Benign (2 rules)
B1: 1 Stand-alone (BA1)
B2: >= 2 Strong benign
Likely Benign (2 rules)
LB1: 1 Strong benign + 1 Supporting benign
LB2: >= 2 Supporting benign
Why the Point System
The original 18 ACMG rules left gaps -- certain evidence combinations had no defined classification. The Bayesian point system fills these gaps while producing identical results for all combinations covered by the original rules. It also handles conflicting evidence more naturally through point summation rather than binary VUS defaults.