Chapter 4: Databases for Clinical Variant Interpretation: ClinVar, ClinGen & HPO
Chapter 4: Databases for Clinical Variant Interpretation: ClinVar, ClinGen & HPO
No clinical geneticist works in isolation. Interpreting genetic variants relies on centralized knowledgebases that aggregate worldwide clinical observations, functional laboratory assays, and population frequencies.
- 0 Stars: Submitted without assertion criteria or single submitter without supporting evidence.
- 1 Star: Criteria provided, single submitter, or conflicting interpretations between labs.
- 2 Stars: Multiple submitters with assertion criteria provided, concordant interpretation.
- 3 Stars: Reviewed by an FDA-recognized ClinGen Variant Curation Expert Panel (VCEP). Gold standard for diagnostic reporting.
- 4 Stars: Established professional practice guideline (e.g. ACMG practice guideline).
1. Population Frequency Filtering with gnomAD
If a variant occurs in the general population at a frequency higher than the disease prevalence in the clinic, it cannot be the cause of a fully penetrant Mendelian disease. ClinGen defines gene-specific maximum credible population allele frequencies (e.g. \(f_{\text{max}} < 0.0001\) for severe dominant pediatric disorders).
2. Phenotype-Driven Prioritization with HPO
The Human Phenotype Ontology (HPO) provides standardized, machine-readable vocabularies for human clinical symptoms (e.g. HP:0001250: Seizure, HP:0001263: Global developmental delay). Tools like Exomiser combine semantic HPO phenotype similarities with variant pathogenicity scores to rank candidate diagnoses.
# Running Exomiser phenotype matching from CLI
java -jar exomiser-cli-13.3.0.jar \\
--analysis-mode PASSENGER \\
--vcf patient_exome.vcf.gz \\
--hpo-ids HP:0001250,HP:0001263,HP:0000252 \\
--output-prefix patient_diagnostic_rankings