Skip to content

Rare variants

Not a medical test — for research and educational use only

Yeliztli analyses consumer genotyping-array data (23andMe / AncestryDNA), which is not a clinical-grade test. Results are not diagnostic, are not clinically validated, and must not be used to make medical decisions. Array data is especially unreliable for rare, disease-causing variants. Always confirm any finding with an accredited clinical laboratory and discuss it with a qualified clinician or genetic counsellor before acting on it. See the Intended use & disclaimers page for the details and the evidence behind this warning.

The Rare Variant Finder is a flexible tool for surfacing the rare and ultra-rare variants you actually carry, with filters you control.

How it runs

The finder runs automatically after annotation as part of the all-modules pipeline. That automatic run is intentionally broad: it stores carried-only variants below the default gnomAD allele-frequency threshold of 1%, using population-max AF when available and global AF as a fallback, plus variants with no gnomAD AF. It also applies the resolved biological sex gate, so incompatible sex-chromosome findings are dropped.

This can produce tens of thousands of low-evidence Rare Variant Finder rows for a typical sample. Most are discovery-context inventory, not diagnoses and not known disease associations. Use the interactive filters, categories, and evidence ratings to narrow the set before reviewing or exporting.

What it looks at

You can set the filters: a gene panel, a gnomAD allele-frequency threshold, whether to include variants with no gnomAD AF, ClinVar significance, predicted consequence type, and zygosity. CADD and REVEL scores are shown in the result details when available.

What you'll see

Findings sorted by clinical relevance, in these categories:

  • ClinVar pathogenic — known Pathogenic/Likely-Pathogenic variants.
  • ClinVar lower-penetrance/risk allele — lower-penetrance/risk-allele findings are stored under a distinct findings category from high-penetrance P/LP variants. The category spans both ClinVar low-penetrance assertions (such as Pathogenic, low penetrance) and risk-allele assertions (such as Established risk allele). A low-penetrance assertion still carries a Pathogenic or Likely pathogenic primary classification; it is held separately because the ClinGen Low Penetrance/Risk Allele Working Group recommends that assertions carrying a penetrance modifier be curated and reported distinctly from standard high-penetrance results.1
  • Ensemble pathogenic — a supporting computational flag based on a strict majority of the prediction axes with data; the reference explains the axes, thresholds, and displayed fraction.
  • Novel — not catalogued in gnomAD, dbSNP, or ClinVar.
  • Rare — other variants passing your frequency filter.

Each carries an evidence rating, rsID, gene, consequence, HGVS, population frequencies, ClinVar details, CADD/REVEL scores, zygosity, and inheritance. You can export results to TSV or VCF.

Good to know

  • Only variants you carry are shown — arrays report a call at every probe regardless of biology, so unscoreable and reference calls are filtered out.
  • Y-chromosome findings are dropped for XX samples.
  • "Novel" requires absence from gnomAD and dbSNP/ClinVar — a gnomAD gap alone isn't enough.
  • A frequency belongs to one specific allele, not to an rsID. Where an rsID covers several alternate alleles and no single frequency can be attributed to your call — because the call carries more than one of them, carries none, or the alleles cannot be matched — no gnomAD frequency is shown rather than another allele's, so the variant is treated as having no gnomAD AF and appears only when you include those.

  1. Schmidt RJ, et al. Recommendations for risk allele evidence curation, classification, and reporting from the ClinGen Low Penetrance/Risk Allele Working Group. Genetics in Medicine. 2024;26(3):101036. PMID:38054408; DOI:10.1016/j.gim.2023.101036 (accessed 2026-07-31). ↩