After donor-level refinement, candidate variants are assigned a final confidence designation based on the level of supporting evidence observed across sequencing technologies, variant callers, and tissues.
The final confidence designation summarizes the validation results obtained during the pipeline and provides an overall assessment of the likelihood that a variant represents a true somatic mosaic event.
Confidence labels are recorded in the FILTER field of the final VCF file.
Variants are classified into three confidence levels:
Variants of the highest confidence.
Variants are labeled HighConf when they satisfy one of the following conditions:
CrossTechCrossTissue and CrossCallerThese variants have independent evidence across multiple sequencing technologies or strong supporting evidence across tissues and variant callers.
Variants of lower confidence that still show partial supporting evidence.
Variants are labeled LowConf when they satisfy one of the following conditions:
CrossTissueCrossCallerThese variants have supporting evidence from either multiple tissues or multiple variant callers but lack cross-technology confirmation.
Variants of the lowest confidence.
Variants are labeled LikelyArtifact when:
CrossTech, CrossTissue, or CrossCaller)These variants pass all upstream filters but lack independent supporting evidence. In most cases they represent sequencing or alignment artifacts, although some may represent true mosaic variants in samples with limited long-read coverage.
The confidence designation provides a simplified summary of the evidence supporting each variant while preserving the detailed annotations generated during the pipeline.
Users may apply additional filtering strategies depending on the desired balance between sensitivity and specificity.
All the relevant code can be accessed in the GitHub repository: