Interpreting Germline and Somatic Variants: Avoiding Oversimplified Rules Based on VAF
Why Is This Important?
After identifying a variant in genomic testing, it is necessary to distinguish whether the variant is present throughout the body or arose specifically in certain tissues or tumors. Germline variants may be present in multiple tissues and germ cells starting from the zygote and can be transmitted to subsequent generations. Somatic variants arise in specific cell lineages post-fertilization and are restricted to their descendant cells.
However, biology does not divide cleanly into these two categories. Variants arising during development can create mosaicism, present only in some normal tissues, and clones acquired in blood cells may appear germline in blood tests. Therefore, one must not definitively determine the origin based solely on data from a single specimen.
Core Concepts
The variant allele fraction (VAF) represents the proportion of sequence reads supporting a variant allele at a given locus. In an ideal diploid normal sample, a heterozygous germline variant may approximate 50%, while a homozygous variant may approach nearly 100%. However, these values represent expectations rather than definitive diagnostic thresholds. Actual VAFs are influenced by sampling variation, mapping bias, copy number alterations, and sample contamination.
VAF in tumor samples is more complex, reflecting the combined effects of tumor cellularity, the proportion of clones harboring the variant, locus-specific amplifications or deletions, loss of heterozygosity, and normal cell infiltration. Identical VAFs can arise from distinct biological scenarios; a high VAF does not necessarily indicate a germline origin, nor does a low VAF definitively signify a sequencing error.
Illustrative Example Using Numerical Reasoning
If a tumor sample harbors a heterozygous variant with normal copy number and the tumor cell fraction is 50%, the expected variant allele frequency (VAF) would be approximately 25%. However, this value increases if the mutant allele is amplified or the wild-type allele is lost. Conversely, the VAF decreases if the mutation is present only in a subset of subclones.
Even when a VAF of approximately 10% is observed in normal blood, it is necessary to distinguish between low-level mosaicism, clonal hematopoiesis, contamination, and technical artifacts. Numerical data narrow the range of hypotheses but do not independently prove the origin.
The Role of Matched Normal Samples
Concurrent analysis of tumor and normal specimens enables the comparison of variants shared between both samples against those enriched in the tumor. However, blood is not always an ideal normal control. In cases where hematologic malignancy, hematopoietic mosaicism, or clonal hematopoiesis is suspected, alternative tissues such as skin fibroblasts may be required.
While familial testing aids in assessing heritability and segregation, it does not exclude all forms of somatic mosaicism in the patient. Selecting appropriate specimens for the specific testing objective is as critical as the analytical algorithms employed.
Depth and Limit of Detection
Read depth increases the opportunity for observation but does not represent overall assay performance. Detecting low allele fractions requires validation of library preparation, unique molecule count, base and mapping quality, strand balance, and position-specific error profiles. Thousands of reads, including PCR duplicates, may not correspond to thousands of independent molecules.
The laboratory defines the limit of detection, sensitivity, specificity, and reproducibility for each variant type and allele fraction. Performance characteristics established for SNVs cannot be directly applied to indels, CNVs, or structural variants.
Common Misconceptions
- VAF and cellular prevalence are not equivalent values.
germlinedoes not imply pathogenicity, nor doessomaticimply an oncogenic driver.- In tumor-only assays, it is essential to distinguish between possible germline findings and confirmed germline results.
- High depth does not automatically resolve incorrect mapping in homologous regions or systematic errors.
Limitations
VAF interpretation requires knowledge of specimen type, collection timing, tumor purity estimation method, local copy number, treatment history, and pipeline version. Without this information, even sophisticated decimal precision yields measurements with substantial uncertainty. Confirmation of clinical origin may require independent normal tissue and a validated germline assay.
Reading in Context
This section connects to the concepts of compound heterozygosity and allele phasing for handling variant allele configurations, as well as the selection among whole-genome sequencing (WGS), whole-exome sequencing (WES), and targeted panels, by comparing testing coverage and blind spots.