Chapter 8: Step 6 — Clinical & Population Repositories: gnomAD, ClinVar, COSMIC & OMIM
Tested Stacks:
GATK 4.6.1
Samtools 1.21
Nextflow 24.10
Seurat v5.1
Python 3.12
Prerequisites: Basic Linux CLI & Sequencing Formats
📚 Chapter 8 of 10
0% Course Completed
Chapter 8: Step 6 — Clinical & Population Repositories: gnomAD, ClinVar, COSMIC & OMIM
Even with precise molecular consequence and in silico scores, a variant cannot be diagnosed in a clinical vacuum. Step 6 of the clinical pipeline interrogates four foundational biomedical databases to establish population frequency, archived clinical consensus, somatic cancer recurrence, and Mendelian inheritance patterns.
Step 6 Technical Specifications
- Core Repositories: gnomAD (Population Controls) • ClinVar (Archived Submissions) • COSMIC (Somatic Hotspots) • OMIM (Gene-Disease Phenotypes).
- Contributed Pipeline Evidence: Control allele frequencies ($AF$), ClinVar review status star tiers, somatic recurrence counts, and disease inheritance modes.
- Clinical Utility: High population control frequency definitively proves a variant is benign, while ClinVar consensus and OMIM validation provide the legal and medical basis for clinical reporting.
The 4 Essential Repositories Master Matrix
| Database Name | Primary Clinical Role | Scale & Coverage | Contributed Pipeline Evidence | ACMG Evidence Triggers |
|---|---|---|---|---|
| gnomAD | Population Control Frequency | >800,000 diverse human exomes & genomes | Determines whether a variant is common or extremely rare in healthy individuals. High frequency (>1% or >5%) proves a variant is benign. | BA1 (>5%), BS1 (>expected), PM2 (absent / ultra-rare ≤0.001%) |
| ClinVar | Archived Clinical Assertions | >3 million clinical variant records from diagnostic labs | Public registry of variants mapped to clinical disease classifications submitted by accredited diagnostic labs (Invitae, GeneDx, Ambry). | PP5 (Reputable pathogenic), BP6 (Reputable benign), Star curation |
| COSMIC | Somatic Cancer Catalogue | >37 million coding mutations across 1.5M tumor samples | Tracks mutation frequencies in tumor tissue specimens to highlight recurrent cancer driver hotspots and therapeutic targets. | Somatic Actionability Tiers (AMP/ASCO/CAP), Hotspot validation |
| OMIM | Gene-Disease Inheritance | >26,000 curated Mendelian entries | Authoritative compendium linking target genes to specific clinical syndromes and inheritance patterns (Autosomal Dominant/Recessive). | Verifies gene-disease validity, establishes inheritance criteria (dominant vs recessive) |
1. gnomAD: Evaluating Population Control Allele Frequencies
The fundamental rule of Mendelian medical genetics is that a variant cannot be more common in the general population than the disease it is claimed to cause. If a genetic disease has a prevalence of 1 in 10,000 individuals, any variant present at 2% in gnomAD is mathematically impossible to be a fully penetrant causal allele.
- Stand-Alone Benign Threshold (BA1): Allele frequency $> 5\%$ in gnomAD across any major continental subpopulation (e.g. European, African, East Asian, South Asian, Latino).
- Rarity for Pathogenic Consideration (PM2_Supporting): Absent from gnomAD, or allele frequency $< 0.001\%$ ($1 imes 10^{-5}$) for dominant diseases.
2. ClinVar: Navigating the 4-Star Review Status Hierarchy
Not all ClinVar records carry equal diagnostic weight. When curating clinical variants, pathologists must verify the submission's review status:
- 0 Stars (No assertion criteria provided): Single laboratory submission without formal scoring documentation. Low evidentiary weight.
- 1 Star (Criteria provided, single submitter OR conflicting interpretations): Multiple submitters disagree (e.g. 2 labs call it VUS, 1 calls it Likely Pathogenic). Requires independent curation.
- 2 Stars (Criteria provided, multiple submitters, no conflicts): Consistent consensus reached across independent accredited diagnostic laboratories. Strong clinical benchmark.
- 3 Stars (Reviewed by an expert panel): Curation performed by a recognized ClinGen Variant Curation Expert Panel (VCEP) (e.g. ClinGen PTEN, CDH1, or Hearing Loss VCEPs). Overrides all 1-star and 2-star records.
- 4 Stars (Practice guideline): Formal ACMG or professional society consensus recommendation.
⚠️
Common Pitfalls & Troubleshooting Checklist
- Java Heap Exhaustion (GATK / Picard): If processes terminate with
java.lang.OutOfMemoryError, set your environment variable explicitly:export _JAVA_OPTIONS="-Xmx16g -Djava.io.tmpdir=/tmp". - Index Timestamp Mismatch: If tools complain that a BAM index is missing or out of date, regenerate with
samtools index -@ 4 sample.bam. - Conda Channel Conflicts: Avoid ABI incompatibilities between bioconda and conda-forge by setting strict priority:
conda config --set channel_priority strict. - Chromosome Naming Contamination: Always verify whether reference FASTA headers use
chr1or1; divergent naming causes silent alignment drops.
Was this guide helpful to your research?
Help us maintain 100% accurate, peer-reviewed bioinformatics standards.