Documentation / Classification / Conflicting Evidence
Conflicting Evidence
The Problem
A variant may trigger both pathogenic criteria (e.g., PM2 + PP3) and benign criteria (e.g., BS1). What classification should it receive? This is one of the most clinically important edge cases in variant interpretation.
Two-Level Approach
Helix Insight handles conflicting evidence at two levels, prioritizing clinical safety.
Level 1: High-Confidence Conflict Safety Check
When pathogenic evidence at Strong or Very Strong level (PVS or PS criteria triggered) directly conflicts with Strong benign evidence (BS criteria triggered), the variant is flagged for manual review regardless of the point total. This prevents automated resolution of genuinely conflicting high-quality evidence.
Level 2: Bayesian Point Summation
For all other conflicts, the Bayesian point system handles them naturally through summation. For example, PM2 (+2) and BS1 (-4) yield a net of -2 = Likely Benign. This is more nuanced than the previous approach (v3.3) which defaulted all conflicts to VUS.
Example
A variant with PM2 (+2) + PP3_Supporting (+1) + PP2 (+1) but also BS1 (-4). Net total: +2 +1 +1 -4 = 0 points = VUS. The point system produces a VUS classification because the strong benign evidence (elevated frequency) substantially counteracts the moderate pathogenic evidence. Under the original ACMG rules, this combination had no defined classification.
BA1 Exception
BA1 is excluded from conflict checks because it operates at a higher priority level. An allele frequency above 5% results in Benign classification regardless of any other evidence, including ClinVar assertions and pathogenic criteria.
Clinical Safety Rationale
In clinical genetics, false positives (reporting a benign variant as pathogenic) can lead to unnecessary invasive procedures, patient anxiety, and cascade testing of family members. When high-quality evidence genuinely conflicts, flagging for manual review is the safest outcome -- it prompts the geneticist to gather additional evidence rather than acting on incomplete information.