Skip to main content
Upload the relevant classification framework and the supporting evidence for a variant. Anara walks through each criterion, states what the evidence supports and what it does not, and returns a classification with the reasoning traceable to your uploaded documents. For clinical geneticists and genomics researchers who need a structured, auditable application of the classification framework.

​
1. Describe the task

ACMG/AMP classification requires working through multiple pathogenicity rule categories and up to 28 individual criteria. Done carefully, it takes 30 to 60 minutes per variant. Done in a rush, it produces classifications that do not hold up. Anara applies the framework from your uploaded guidance documents to the variant and evidence you provide. It works through each criterion, states what the evidence supports and what it does not, and returns a classification with the reasoning at each step. If evidence is insufficient for a criterion, it says so. Here a clinical geneticist at a diagnostic laboratory is classifying a CFTR variant for a patient with atypical cystic fibrosis phenotype, working through the ACMG/AMP framework with the published CFTR-specific classification guidance and population frequency data from gnomAD. 
I need to classify this CFTR variant: c.1521_1523delCTT (p.Phe508del). I have uploaded: the ACMG/AMP variant classification guidelines (Richards et al. 2015), the CFTR2 program classification guidance, and the relevant population frequency data from gnomAD v3.

Apply the ACMG/AMP criteria to this variant using only the uploaded documents and the data I have provided. Walk through each criterion category:
- PVS1 (null variant)
- PS1-PS4 (strong pathogenic)
- PM1-PM6 (moderate pathogenic)
- PP1-PP5 (supporting pathogenic)
- BA1, BS1-BS4, BP1-BP7 (benign criteria)

For each criterion, state whether it applies, does not apply, or whether the evidence is insufficient to evaluate. Give the reason for each decision with reference to the specific guideline text. State the final classification and the combination of criteria that produced it.

​
2. Give Anara context

Required context
  • The classification framework as a PDF upload: ACMG/AMP criteria, InSiGHT position statement, disease-specific classification guidance, or the equivalent framework for your application.
  • The variant data: genomic coordinates, transcript, protein change, and any available functional, population frequency, or segregation data.
Optional context
  • Published case reports or functional studies relevant to this specific variant or the same codon. Anara incorporates them into the relevant criteria evaluation.
  • Prior classifications of the same variant or closely related variants from ClinVar or your institution. Conflicting prior classifications are worth flagging in the reasoning.

​
3. What Anara creates

A criterion-by-criterion classification walk-through with the final ACMG/AMP classification and the combination of criteria that produced it. Each decision cites the guideline text and states whether the evidence meets the criterion, does not meet it, or is insufficient. The output is the documentation of reasoning, structured to support institutional review and match the documentation format required for reporting.

​
4. Follow-up prompts

​
Apply pharmacogenomics classification guidance

When the variant affects drug metabolism rather than disease risk and requires a different framework. 
I need to apply CPIC pharmacogenomics classification to this SLCO1B1 variant in the context of simvastatin prescribing. I have uploaded the CPIC guideline for SLCO1B1 and simvastatin. Walk through the evidence and return the CPIC classification with the relevant activity score and prescribing implication.

​
Flag criteria that need additional evidence

When the classification is uncertain and you want to know what data would resolve it. 
For the criteria you rated as insufficient evidence, tell me: what specific type of data would be needed to evaluate each criterion? What is the most practical source for each data type, and what change to the classification would each positive result produce?
When a functionally similar variant has an established classification and you want to assess whether the reasoning transfers. 
The variant c.1521_1523delCTT is a deletion of phenylalanine at position 508. Search my library for any prior classification documents that address this position or nearby in-frame deletions in CFTR. Does any prior reasoning for an adjacent variant carry over to this one under the PS1 or PM5 criteria?

​
5. Tricks, tips, and troubleshooting

​
How you word your prompt shapes what you get

Provide the variant in both HGVS notation and a plain description of the change type (in-frame deletion, missense, splice variant). This reduces errors in the PVS1 evaluation. Listing the evidence you have before asking Anara to classify tells it which criteria are evaluable and which will default to insufficient evidence, which is faster than working through every criterion in sequence.

​
Check the output against your own understanding

Anara applies the framework from the documents you upload. It does not access ClinVar, functional databases, or literature not in your library. If a relevant functional study or known database classification has not been uploaded, it is not in the assessment. For variants where the classification determines a clinical action, laboratory director review remains required. Anara is decision support for the workflow, not the classification itself.

​
What to do with the output next

Use the criterion-by-criterion reasoning as the basis for CLIA variant classification documentation. Where criteria are marked insufficient evidence, document that status and flag the variant for reclassification if new data emerges. For variants of uncertain significance, the follow-up prompt on additional evidence types identifies whether a reclassification pathway exists or whether the uncertainty is stable.