An end-to-end in silico framework for truth-based validation and regression testing of clinical genome sequencing pipelines, from simulated reads to reportable findings

An end-to-end in silico framework for clinical genome pipelines achieved 98.3% report inclusion of 115/117 pathogenic and likely pathogenic variants in GIAB HG002.

• Why it matters: Pipeline updates require proportionate revalidation, but existing in silico guidance rarely tests variant interpretation through to reportable decisions, leaving report specificity unmeasured.
• What they did: The framework plants ClinGen-curated pathogenic/benign variants (and optional synthetic ones) into the GIAB HG002 germline background, simulates reads, runs secondary/tertiary stages as black boxes, and compares truth at call-set and report levels.
• The result: For nf-core/sarek with DeepVariant and ACMG/AMP classifier BIAS-2015 across five test sets (243 variants), report sensitivity was 98.3% and it withheld 106/126 non-reportable variants (84.1%), with false reports mainly linked to diploid calling of male chromosome X.

AI summary of the abstract · may contain errors · not medical advice · read the paper.