Biomarker evidence that holds up to scrutiny.
95% of candidate biomarkers fail to demonstrate clinical utility. Most validation pipelines pull from a single source, miss contradictions between studies, and don't reconcile AMP, OncoKB, and ESCAT tiers. Skippy Biomarker fuses 20+ sources with Bayesian aggregation, flags contradictions with fragility scoring, and generates FDA RBM-structured evidence packages.
What a response looks like
Biomarker validation with Bayesian confidence, multi-framework tier emission, and contradiction detection — all three in one API.
Pembrolizumab (PD-1/PD-L1 inhibitor) — FDA-approved first-line for TPS ≥50%
Simulated output representative of real API responses. Bayesian Beta-Binomial fusion across 20+ sources. AMP/ASCO/CAP, OncoKB, ESCAT, ACMG tiers.
Six capabilities. One API.
Biomarker validation
Validate any biomarker-disease pair across 20+ curated sources — ClinVar, CIViC, OpenTargets, FDA Labels, ESMO, NCCN, and clinical trial registries. Returns Bayesian confidence with credible interval, role classification (predictive / prognostic / diagnostic), and retraction check.
Multi-framework tier emission
Emit AMP/ASCO/CAP (IA–IVD), OncoKB (Level 1–4), and ESCAT (I-A through IV) tier assignments simultaneously for somatic biomarkers. ACMG Pathogenic/LP/VUS germline classification available for dual-role variants. All three frameworks reconciled in a single call.
Contradiction detection
Identifies conflicting claims across sources — endpoint mismatch, curation cutoff discrepancies, cohort heterogeneity. Returns fragility index (studies-to-overturn), leave-one-out sensitivity, and root-cause hypothesis. Confidence is adjusted downward when contradictions are unresolved.
Regulatory evidence package
Generate FDA RBM qualification-ready or EMA Article 10-scoped evidence packages: executive summary, analytical validation, clinical validation, clinical utility, and references. Structured to LOI → Qualification Plan → Data Submission workflow. Weeks instead of months.
Continuous evidence monitoring
Continuous watch on ClinVar, PubMed, and clinical trial registries for new evidence. When incoming data warrants tier reclassification or contradicts the current confidence, an alert fires automatically — without waiting for annual review.
Ancestry equity audit
Flag biomarkers where clinical evidence derives predominantly from a single ancestry group — a systematic source of validation bias. Returns per-cohort coverage gaps and confidence impact, with guidance on where additional trial data is needed for equitable evidence generation.
AMP/ASCO/CAP, OncoKB, ESCAT, and ACMG — reconciled.
Three independent tiering systems use different evidence standards and terminology. Skippy Biomarker emits all frameworks simultaneously — so a single API call tells you where a biomarker stands across every guideline body a regulator or payer will ask about.
FDA-approved therapy available for this biomarker in the tested tumor type. Highest actionability.
Well-powered clinical trial (not yet FDA-approved) demonstrates predictive value with durable response.
Small study, case reports, or preclinical evidence. Biological plausibility established; clinical validation incomplete.
Level 1: FDA-recognized biomarker with approved therapy. Level 2: Standard-of-care biomarker per major guidelines.
ESCAT I-A: Prospective randomized trial in the target population. I-B: Prospective non-randomized or retrospective with strong statistical power.
Available for dual-role variants. Tavtigian Bayesian posterior — Pathogenic ≥ 0.99, Likely Pathogenic ≥ 0.90. Supports companion diagnostic CDx design.
CIViC and OncoKB can disagree on the same biomarker. ASCO GI 2023 and NEJM 2024 can reach opposite conclusions on the same drug-disease pair. Most validation tools don't tell you this.
Skippy Biomarker detects conflicting claims between sources, identifies the root cause (endpoint mismatch, curation cutoff, cohort heterogeneity), and adjusts Bayesian confidence downward when contradictions are unresolved. Fragility index tells you how many studies it would take to overturn the current evidence — a fragility index of 2 is a different clinical reality than a fragility index of 20.
“Validation programs that rely on a single curated database inherit that database's curation cutoff date, inclusion criteria, and ancestral cohort distribution as silent assumptions.”
Skippy Biomarker fuses 20+ sources with TruthFinder source-reputation weighting — sources with retraction histories or curation lag are down-weighted automatically. Ancestry equity audits surface evidence gaps for non-European populations before they become regulatory findings.
Pharma, labs, biotech, and academia
FDA RBM, EMA Article 10, and CLIA/CAP
Evidence packages structured to FDA Biomarker Qualification Program requirements — Letter of Intent, Qualification Plan, Data Submission, and Qualification Opinion stages. EMA Biomarker Qualification Opinion (Article 10 protocol) output also available.
CLIA/CAP analytical validation standards supported for clinical laboratory accreditation. AMP/ASCO/CAP somatic tiering with full evidence provenance meets tumor board reporting standards.
HIPAA-ready with BAA available. Every assessment produces an immutable audit record with all source attributions, Bayesian parameters, tier assignments, and contradiction flags.
Often deployed together
Germline variant interpretation with ACMG/AMP Bayesian classification. Dual-role variants — both germline ACMG pathogenicity and somatic AMP/ASCO/CAP tiering — available in the same API call.
Biomarker-guided drug repurposing: genetic evidence from GWAS and ClinVar feeds the Swanson ABC pipeline. Validated biomarkers from Skippy Biomarker directly inform repurposing candidate scoring.
CDx companion to PGx testing — biomarker tier status informs whether a gene-drug interaction rises to companion diagnostic or pharmacogenomic advisory level.
Skippy Biomarker is a decision-support tool for qualified healthcare and research professionals. It does not replace pathologist review, laboratory director sign-out, clinical validation studies, or physician interpretation. All biomarker tier assignments and confidence scores should be confirmed against current institutional guidelines, applicable regulatory approvals, and patient-specific clinical context before use in clinical decisions. HIPAA-ready; BAA available.
See Biomarker in your validation program
We work with pharma R&D, clinical labs, and biotech companies. Let's talk about your biomarker validation problem.