What is VAF (Variant Allele Frequency)?
VAF is the fraction or percentage of sequencing reads that contain a specific genetic variant at a given genomic position, serving as a quantitative measure of the proportion of mutated DNA in a sequenced sample 1.
Core Definition and Measurement
VAF represents a fundamental metric in molecular diagnostics that quantifies mutation burden. Specifically, it is calculated as the number of reads supporting a variant divided by the total read depth at that locus 1, 2. For example, if 15 out of 100 sequencing reads show a mutation, the VAF is 15% or 0.15.
Clinical Context and Interpretation
In Tumor Tissue Analysis
In cancer tissue samples, VAF reflects:
- Tumor clonality: Higher VAF typically indicates the mutation is present in most or all cancer cells (clonal), while lower VAF suggests subclonal populations 1
- Tumor purity and ploidy: The relationship between VAF and actual cellular prevalence depends on tumor purity (fraction of cancer cells) and copy number alterations 1
- Intratumoral heterogeneity: VAF patterns reveal the evolutionary architecture of the tumor 3, 4
A critical caveat: VAF alone does not equal cancer cell fraction (CCF). A 50% VAF in a pure diploid tumor with a heterozygous mutation represents 100% of cancer cells, but the same VAF in a sample with 50% tumor purity could represent different scenarios depending on copy number 1.
In Liquid Biopsy (ctDNA Analysis)
For circulating tumor DNA, VAF interpretation differs fundamentally 5:
- High plasma VAF indicates elevated tumor burden and is associated with worse outcomes (opposite to tissue interpretation) 3
- VAF dynamics during treatment predict response and can detect resistance early 3, 6
- Low VAF (<1%) may represent clonal hematopoiesis rather than tumor-derived variants, particularly in genes like DNMT3A, TET2, ASXL1, and JAK2 5
Technical Thresholds and Reporting Standards
Detection Limits
Multiple guidelines establish VAF thresholds for clinical reporting 2:
- Sanger sequencing: Detects variants ≥10% VAF reliably
- NGS minimum reporting threshold: 10% VAF for clinical decision-making 2
- Variants between 5-10% VAF: May be reported with explicit disclaimer that clinical impact is unproven 2
- Limit of Blank (LoB): Variants at or below this level should not be reported 5
- Between LoB and Limit of Detection (LoD): Should be labeled "equivocal" with uncertainty disclaimer 5
Quality Requirements
For reliable VAF determination 2:
- Minimum 100× read depth per position
- At least 10 variant-supporting reads
- Bidirectional strand support (forward and reverse)
- No bias in read position, base quality, or mapping quality
Clinical Applications
Prognostic and Predictive Value
VAF serves multiple clinical roles 3, 4:
- Risk stratification: Baseline VAF enables patient stratification in targeted therapy
- Treatment monitoring: Serial VAF measurements track response and detect resistance 6
- Resistance mechanisms: In EGFR-mutant NSCLC, the T790M/activating mutation VAF ratio predicts osimertinib efficacy 3
Germline vs. Somatic Distinction
VAF helps distinguish variant origin 7, 8:
- ~50% VAF: Suggests heterozygous germline variant
- ~100% VAF: Suggests homozygous germline or loss of heterozygosity
- Variable VAF (<50%): Typically somatic or mosaic
- Important limitation: VAF alone is insufficient—tumor purity and copy number must be considered 7
Key Pitfalls to Avoid
- Do not equate VAF with tumor cell percentage without accounting for purity and ploidy 1
- Do not assume low VAF variants are clinically irrelevant—they may represent actionable subclones 2
- Do not interpret plasma VAF the same as tissue VAF—high plasma VAF indicates poor prognosis while high tissue VAF may indicate favorable response 3
- Do not ignore clonal hematopoiesis in liquid biopsy, especially for variants <1% VAF in CHIP-associated genes 5
- Do not use VAF from ctDNA to determine germline origin—it is not informative for this purpose 7