ACMG Classification¶

AIVA includes an interactive ACMG/AMP variant classifier that lets you apply evidence-based criteria to individual variants. The classifier automatically calculates the five-tier pathogenicity classification based on the criteria you select.

For details on the ACMG/AMP framework and how to use the public classifier, see Using the Variant Classifier.

Using the Classifier¶

Step 1: Open the classifier¶

1. Open a sample in the Data Table or Tertiary Analysis view.
2. Locate the variant you want to classify.
3. Click the ACMG Classification action on the variant row.
4. The classification panel opens, displaying all 28 criteria organized by category.

Step 2: Select applicable criteria¶

1. Review each criterion and determine whether it applies to the variant based on available evidence.
2. Check the criteria that are met. You can select any combination of pathogenic and benign criteria.
3. For each selected criterion, you may add a note documenting the specific evidence (e.g., "gnomAD AF = 0.0001, absent in controls").

Step 3: Review the auto-calculated classification¶

As you select criteria, the classifier automatically calculates the resulting classification using the ACMG/AMP combining rules. The classification updates in real time as you add or remove criteria.

Step 4: Save the classification¶

1. Review the selected criteria and the auto-calculated classification.
2. Click Save to store the classification.
3. The classification is associated with the variant and visible in the data table and to all project collaborators.

Evidence Tracking¶

Each ACMG classification in AIVA includes:

• The selected criteria with any associated notes.
• The auto-calculated classification tier.
• The user who performed the classification.
• The timestamp of the classification.
• A history of changes if the classification is updated over time.

This evidence trail supports clinical documentation and audit requirements.

Collaborative Classification¶

When a sample belongs to a project:

• Editors and Owners can create and modify ACMG classifications.
• Viewers can view existing classifications but cannot modify them.
• Multiple team members can independently classify the same variant, allowing for comparison and consensus building.

Tips¶

• Document your reasoning: Add notes to each selected criterion explaining the specific evidence.
• Start with population frequency: Check gnomAD allele frequency first. If AF > 5%, the variant is classified as Benign (BA1). This eliminates many variants quickly.
• Revisit VUS variants: Variants classified as VUS should be periodically re-evaluated as new evidence becomes available.