Bio Clinical Databases Acmg Classification

Version Compatibility

Reference examples tested with: requests 2.31+, pandas 2.2+, AutoPVS1 (Xiang 2020), InterVar 2.2+, GeneBe 1.0+ (Stawiński 2024 Clin Genet). ACMG/AMP Bayesian point system is Tavtigian 2018 Genet Med / 2020 Hum Mutat. Pejaver 2022 AJHG PP3/BP4 calibrated thresholds. ClinGen Splicing Subgroup 2023 (Walker AJHG). v3.2 ACMG SF list (Miller 2023). The ACMG 2.0 framework is in development as of May 2026; not yet published.

Before using code patterns, verify installed versions match. If versions differ:

• Python: pip show <package> then help(module.function) to check signatures
• CLI: <tool> --version

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying. VCEP-specific CSpec rules override default ACMG application; the authoritative directory is https://cspec.genome.network/cspec/ui/svi/all.

ACMG/AMP Variant Classification Framework

'Classify this variant per ACMG/AMP' -> Apply 28-criterion framework using Tavtigian point system; gate on ClinGen SVI specifications and VCEP-specific overrides; assign P / LP / VUS / LB / B classification with evidence trail.

• Python (automated): GeneBe API https://api.genebe.net/cloud/api-public/v1/variant
• Python (rule-based): InterVar -> python InterVar.py -i input.vcf -b hg38 --table_annovar table_annovar.pl
• Web tools: VarSome (commercial), Franklin/Genoox (commercial), ClinGen VCI (gold standard for SVI)
• Citation: Richards 2015 Genet Med 17:405 (original framework); Tavtigian 2020 Hum Mutat 41:1734 (point system)

The Tavtigian Bayesian Point System: The Engine Inside All Modern Classifiers

Richards 2015 specified 28 criteria with strength labels (Supporting / Moderate / Strong / Very Strong); combination rules produced P / LP / VUS / LB / B. Tavtigian 2018/2020 demonstrated this framework is mathematically a Bayesian classifier and proposed the naturally-scaled point system that every modern automated classifier implements:

Benign codes are negative-signed. Final classification:

InterVar / GeneBe / VarSome / Franklin all implement Tavtigian point summation under the hood. Combinations never appearing in the 2015 combining rules (e.g., PVS1_VeryStrong + PM2_Supporting -> LP) emerge naturally from point arithmetic.

PVS1 Decision Tree (Abou Tayoun 2018 Hum Mutat 39:1517)

PVS1 is the most consequential code: pathogenic Very Strong (8 points) for predicted loss-of-function in a gene where LoF is established disease mechanism. The 2018 decision tree refined PVS1 from a binary into a graded code based on:

1. Variant type: nonsense / frameshift / canonical +-1,2 splice / initiation codon / single-exon deletion / multi-exon deletion.
2. NMD prediction: variant in 5'-most exon OR >50bp upstream of last exon-exon junction -> NMD-triggered. Else truncated protein.
3. Critical region: removal of >10% of coding sequence OR removal of a critical functional domain.
4. Alternative isoform: does the variant affect a transcript expressed in disease-relevant tissue?

Output strengths:

Subsumption rule (Abou Tayoun 2018): PVS1 + PP3 -> only PVS1 counts (PP3 is subsumed). Same for PVS1 + PM4.

>15 VCEP-specific PVS1 trees exist as of 2024 (CDH1, ENIGMA BRCA1/2, FH LDLR/APOB/PCSK9, InSiGHT MMR, RASopathies, hearing loss, hypertrophic cardiomyopathy, Rett/Angelman, etc.). The automated implementation is AutoPVS1 (Xiang 2020).

Pejaver 2022 PP3/BP4 Calibrated Thresholds (the load-bearing 2024+ calibration)

Pejaver 2022 AJHG 109:2163 Bayesian-calibrated 13 missense predictors to PP3/BP4 strength levels using ClinVar P/B variants with leave-one-gene-out cross-validation.

The numbers to memorize: REVEL >= 0.932 = PP3_Strong; REVEL <= 0.016 = BP4_Strong (<= 0.003 BP4_VeryStrong); the 0.290-0.644 band is indeterminate (no criterion).

AlphaMissense calibration (Bergquist 2025 Genet Med 27:101402, ClinGen SVI; originally Bergquist et al. bioRxiv 2024.09.17): this ClinGen SVI calibration extends the graded PP3/BP4 options to AlphaMissense, reaching PP3_Strong and at least BP4_Moderate. Critical: the developer-recommended 0.564 threshold is NOT the calibrated PP3 threshold; verify the current ClinGen SVI recommendation for the exact score cutoffs before applying.

Do not stack predictors. REVEL, BayesDel, VEST4 share ClinVar/HGMD training data; using REVEL >= 0.773 AND BayesDel >= 0.27 to claim "two independent moderate hits" is double-counting. Pejaver 2022 explicitly recommends using ONE predictor per variant.

PM2_Supporting (ClinGen SVI 2020)

The original PM2 ("absent from controls") was over-weighted. SVI 2020 downgraded to PM2_Supporting (1 point, not 2). Mechanism: most rare variants are benign. Empirical recalibration showed ~6 variants per gene downgrade from LP to VUS when PM2 -> Supporting. Many 2017-2019 LP curations require re-classification post-SVI 2020 update.

PS3/BS3 Functional Evidence (Brnich 2020 Genome Med 12:3)

OddsPath framework; the four-step SOP:

1. Define disease mechanism for the gene.
2. Evaluate assay class (e.g., MAVE, biochemical, animal model).
3. Evaluate specific assay instance (controls, replicate consistency).
4. Apply per-variant.

OddsPath calibration mapping to ACMG strengths:

MAVEdb deep-mutational scans with >=11 controls (>=5 P/LP + >=5 B/LB) can yield up to PS3_Strong/BS3_Strong via OddsPath calibration. This is the entry point for MAVE/saturation-mutagenesis evidence into ACMG.

Default-Strong PS3 application is increasingly over-strengthening without OddsPath calibration; ClinGen SVI recommends moving toward PS3_Moderate as default unless OddsPath > 4.3.

ClinGen SVI Splicing Subgroup 2023 (Walker AJHG 110:1046)

SpliceAI is the recommended primary splicing tool. Calibrated thresholds:

SpliceAI prediction alone does NOT reach PP3_Strong; strength escalation requires experimental/RNA splicing evidence (PS3) or the repurposed PVS1 route.

SpliceVault / 300K-RNA (Dawes 2023 Nat Genet 55:324): does NOT predict whether a variant is splice-altering; predicts WHAT the aberrant transcript will be (which exon skips, which cryptic site activates). 96% sensitivity for exon-skipping; 86% for cryptic site activation in 140 clinical RNA-tested cases. Critical for PVS1 application to splice variants because PVS1 depends on whether the aberrant transcript triggers NMD.

Pangolin (Zeng 2022 Genome Biol 23:103): SpliceAI improvement for cryptic donor sites; not yet ClinGen-endorsed but increasingly used as tiebreaker.

BS1 / BA1 (Whiffin Max-Credible-AF)

BA1 default: AF > 5% in non-bottleneck group per ClinGen SVI; VCEP-specific overrides (Hearing Loss VCEP uses 0.5% AR).

BS1 gene-specific: (prevalence x heterogeneity x allelic-contribution) / (penetrance x 2) from Whiffin 2017 Genet Med 19:1151. Compare against gnomAD grpmax_faf95.

See clinical-databases/gnomad-frequencies for FAF95 details.

ClinGen VCEP CSpec Hierarchy

ClinGen VCEP CSpec authoritative registry: https://cspec.genome.network/cspec/ui/svi/all. ~80-90 VCEPs as of 2025. Examples:

• Hearing Loss VCEP: PM2 -> supporting default; PS3 thresholds upgraded for OTOF; BA1 lowered to 0.5% AR.
• ENIGMA BRCA1/2 VCEP: gene-specific PVS1 trees with NMD escape rules; PS4 case-control thresholds.
• Inherited Cardiac Conditions VCEP: gene-specific PS4 (5+ unrelated probands for PS4_Supporting).

Apply VCEP CSpec when one exists. Generic ACMG with no VCEP awareness is unreliable for many genes.

Cancer Somatic Framework (Li 2017 J Mol Diagn 19:4)

AMP/ASCO/CAP somatic variant interpretation; four tiers:

Knowledgebases: OncoKB (MSKCC; Chakravarty 2017), CIViC (Griffith 2017 Nat Genet 49:170), CGI (Tamborero 2018), JAX-CKB, COSMIC. OncoKB Levels (1-4 therapeutic) map to AMP tiers loosely.

The Variant Interpretation for Cancer Consortium (VICC) Meta-Knowledgebase standards (2024-2025) harmonize across knowledgebases. ClinGen Somatic VCEPs are emerging (started 2022).

Decision Tree by Variant Type

Standard Workflow: ACMG Classification

Goal: Apply ACMG/AMP framework to a candidate variant with proper SVI specifications and VCEP overrides.

Approach: Pull aggregated evidence; apply Pejaver-calibrated in-silico thresholds; check VCEP-specific CSpec; sum Tavtigian points.

import requests
import pandas as pd


# Pejaver 2022 calibrated REVEL thresholds (one-predictor rule applies)
REVEL_THRESHOLDS = {
    'BP4_VeryStrong': (-float('inf'), 0.003),
    'BP4_Strong': (0.003, 0.016),
    'BP4_Moderate': (0.016, 0.183),
    'BP4_Supporting': (0.183, 0.290),
    # (0.290, 0.644) = indeterminate zone, no criterion applied
    'PP3_Supporting': (0.644, 0.773),
    'PP3_Moderate': (0.773, 0.932),
    'PP3_Strong': (0.932, float('inf'))
}

# Tavtigian point assignments (Tavtigian 2020 Hum Mutat)
STRENGTH_POINTS = {
    'PVS1_VeryStrong': 8, 'PVS1_Strong': 4, 'PVS1_Moderate': 2, 'PVS1_Supporting': 1,
    'PS1': 4, 'PS2': 4, 'PS3': 4, 'PS3_Moderate': 2, 'PS3_Supporting': 1, 'PS4': 4,
    'PM1': 2, 'PM2_Supporting': 1, 'PM3': 2, 'PM3_Strong': 4, 'PM3_VeryStrong': 8,
    'PM4': 2, 'PM5': 2, 'PM6': 2,
    'PP1': 1, 'PP1_Moderate': 2, 'PP1_Strong': 4,
    'PP2': 1, 'PP3_Supporting': 1, 'PP3_Moderate': 2, 'PP3_Strong': 4, 'PP4': 1, 'PP5': 1,
    # Benign codes (negative)
    'BA1': -100,  # Standalone benign
    'BS1': -4, 'BS2': -4, 'BS3': -4, 'BS3_Moderate': -2, 'BS3_Supporting': -1, 'BS4': -4,
    'BP1': -1, 'BP2': -1, 'BP3': -1,
    'BP4_Supporting': -1, 'BP4_Moderate': -2, 'BP4_Strong': -4, 'BP4_VeryStrong': -8,
    'BP5': -1, 'BP6': -1, 'BP7': -1
}


def classify_revel_pp3_bp4(revel_score):
    '''Map REVEL score to PP3/BP4 strength per Pejaver 2022.'''
    if revel_score is None:
        return None
    for code, (lo, hi) in REVEL_THRESHOLDS.items():
        if lo <= revel_score < hi:
            return code
    return None


def classify_alphamissense_supporting_only(am_score):
    '''AlphaMissense is currently supporting-only; ClinGen has not endorsed PP3 calibration.

    Cheng 2023 developer threshold 0.564 is NOT the Pejaver-style PP3 calibration.
    '''
    if am_score is None:
        return None
    if am_score >= 0.7:
        return 'PP3_Supporting'   # Tentative; ClinGen not endorsed
    if am_score <= 0.2:
        return 'BP4_Supporting'   # Tentative
    return None


def spliceai_to_acmg(ds_max):
    '''Walker 2023 SVI Splicing Subgroup framework.

    Computational SpliceAI codes are applied at Supporting weight.
    SpliceAI >= 0.20 -> PP3_Supporting (minimum for ANY splicing PP3).
    SpliceAI <= 0.1 -> BP4_Supporting.
    Prediction alone does not reach PP3_Strong; escalation needs RNA/experimental evidence.
    '''
    if ds_max is None:
        return None
    if ds_max >= 0.20:
        return 'PP3_Supporting'
    if ds_max <= 0.1:
        return 'BP4_Supporting'
    return None


def tavtigian_classify(criteria_assigned):
    '''Sum Tavtigian points and classify P / LP / VUS / LB / B.

    criteria_assigned: list of criterion strings (e.g., ['PVS1_VeryStrong', 'PM2_Supporting'])
    '''
    points = sum(STRENGTH_POINTS.get(c, 0) for c in criteria_assigned)
    if any(c == 'BA1' for c in criteria_assigned):
        return {'classification': 'Benign', 'points': points, 'rationale': 'BA1 standalone'}
    if points >= 10:
        category = 'Pathogenic'
    elif points >= 6:
        category = 'Likely Pathogenic'
    elif points >= 0:
        category = 'VUS'
    elif points >= -6:
        category = 'Likely Benign'
    else:
        category = 'Benign'
    return {'classification': category, 'points': points, 'criteria': criteria_assigned}


def genebe_classify(hgvs):
    '''Query GeneBe API (Stawiński 2024) for automated ACMG classification.

    GeneBe is open-source, Tavtigian-point-system-based, and performs comparably to
    VarSome (which is commercial, 82% ACMG criteria auto-application rate).
    '''
    r = requests.get(f'https://api.genebe.net/cloud/api-public/v1/variant',
                     params={'variant': hgvs, 'genome': 'hg38'},
                     timeout=30)
    r.raise_for_status()
    return r.json()


def whiffin_max_credible_af(prevalence, max_allelic_contribution=1.0,
                              max_genetic_contribution=1.0, penetrance=1.0):
    '''Compute gene-specific BS1 max-credible-AF (Whiffin 2017 Genet Med).

    Returns: max-credible per-allele frequency under dominant inheritance.
    For autosomal recessive, transform appropriately.
    '''
    return (prevalence * max_genetic_contribution * max_allelic_contribution) / (penetrance * 2)


def apply_bs1_ba1(grpmax_faf95, max_credible_af, ba1_threshold=0.05):
    '''Apply ClinGen SVI BS1/BA1 from gnomAD grpmax FAF95.'''
    if grpmax_faf95 is None or grpmax_faf95 == 0.0:
        return 'PM2_Supporting'
    if grpmax_faf95 > ba1_threshold:
        return 'BA1'
    if grpmax_faf95 > max_credible_af:
        return 'BS1'
    return None

Per-Operation Failure Modes

1. Stacking REVEL + BayesDel + VEST4 as independent evidence

• Trigger: Apply PP3 from multiple predictors.
• Mechanism: Predictors share training data; double-counting.
• Symptom: Inflated PP3 strength; over-classified LP.
• Fix: Use ONE predictor per variant (Pejaver 2022).

2. AlphaMissense PP3_Strong with developer 0.564 threshold

• Trigger: Apply AlphaMissense >0.564 -> PP3_Strong.
• Mechanism: 0.564 is the developer-recommended likely-pathogenic threshold, NOT a calibrated PP3 cutoff; use the ClinGen SVI calibration (Bergquist 2025) score thresholds instead.
• Symptom: Over-application of PP3.
• Fix: Use AlphaMissense as supporting evidence only; defer to Pejaver-calibrated REVEL.

3. PVS1 applied to nonsense variant in GoF gene

• Trigger: Nonsense variant in SCN5A reported as PVS1 for LQT3.
• Mechanism: SCN5A has both LoF (Brugada) and GoF (LQT3) mechanisms. PVS1 should NOT apply if LoF is not the established mechanism.
• Symptom: Wrong classification; clinical action mis-directed.
• Fix: Check ClinGen gene-disease mechanism; apply PVS1 only when LoF is established.

4. Generic ACMG instead of VCEP CSpec

• Trigger: Apply default ACMG to a variant in a gene with VCEP-specific CSpec.
• Mechanism: VCEP CSpec overrides for the gene; e.g., Hearing Loss VCEP PM2 default = supporting, BA1 = 0.5%.
• Symptom: Wrong strength applied; misclassification.
• Fix: Check https://cspec.genome.network/cspec/ui/svi/all for active VCEP; apply gene-specific CSpec.

5. PM2 at Moderate (pre-2020 SVI)

• Trigger: Apply PM2 = Moderate (2 points) per 2015 rules.
• Mechanism: SVI 2020 downgraded to PM2_Supporting (1 point).
• Symptom: ~6 variants per gene over-strengthened LP.
• Fix: Use PM2_Supporting per current SVI.

6. PS3 default Strong without OddsPath

• Trigger: Apply PS3 = Strong without OddsPath calibration.
• Mechanism: Default PS3 = Strong over-strengthens; Brnich 2020 SOP requires OddsPath > 4.3 for Strong.
• Symptom: PS3-driven over-classification.
• Fix: Apply Brnich 2020 four-step OddsPath; default move to PS3_Moderate without OddsPath > 4.3.

7. Synonymous treated as no impact

• Trigger: Filter out synonymous variants from classification pipeline.
• Mechanism: Synonymous can disrupt splicing; SpliceAI captures this.
• Symptom: Pathogenic splice-disrupting synonymous missed.
• Fix: Always run SpliceAI on synonymous variants in disease genes; PP3_Supporting if DS_max >= 0.2.

8. ClinVar P + ClinGen Limited validity

• Trigger: Report variant P in gene with Limited gene-disease validity.
• Mechanism: ClinVar P is variant-level; gene-disease validity is upstream.
• Symptom: Mis-attribution to a non-disease gene.
• Fix: Apply ClinGen gene-disease validity gate (Moderate+ for clinical action); for Limited genes, require VCEP curation.

9. Variant on wrong transcript

• Trigger: HGVS-c on alt transcript; functional impact different on MANE Select.
• Mechanism: Tissue-specific isoform considerations; MANE Select 2024+ is clinical standard.
• Symptom: Wrong consequence prediction.
• Fix: Re-evaluate on MANE Select transcript; cross-check with VEP --mane_select.

Reconciliation: When Tools Disagree

Quantitative Thresholds and Conventions

Common Errors

Anticipated Reviewer Pushback

References

• Richards S et al. 2015. Standards and guidelines for the interpretation of sequence variants. Genet Med 17:405. (Original ACMG/AMP)
• Tavtigian SV et al. 2018. Modeling the ACMG/AMP variant classification guidelines as a Bayesian classification framework. Genet Med 20:1054.
• Tavtigian SV et al. 2020. Fitting a naturally scaled point system to the ACMG/AMP variant classification guidelines. Hum Mutat 41:1734.
• Abou Tayoun AN et al. 2018. Recommendations for interpreting the loss of function PVS1 ACMG/AMP variant criterion. Hum Mutat 39:1517.
• Pejaver V et al. 2022. Calibration of computational tools for missense variant pathogenicity classification. Am J Hum Genet 109:2163.
• Bergquist T et al. 2025. Calibration of additional computational tools expands ClinGen recommendation options for variant classification with PP3/BP4 criteria. Genet Med 27:101402.
• Brnich SE et al. 2020. Recommendations for application of the functional evidence PS3/BS3 criterion using the ACMG/AMP sequence variant interpretation framework. Genome Med 12:3.
• Walker LC et al. 2023. ClinGen SVI Splicing Subgroup recommendations. Am J Hum Genet 110:1046.
• Biesecker LG et al. 2024. ClinGen guidance for use of the PP1/BS4 co-segregation and PP4 phenotype specificity criteria for sequence variant pathogenicity classification. Am J Hum Genet 111:24. (PP1/BS4 co-segregation)
• Cheng J et al. 2023. Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science 381:eadg7492.
• Jaganathan K et al. 2019. Predicting splicing from primary sequence with deep learning. Cell 176:535. (SpliceAI)
• Zeng T et al. 2022. Predicting RNA splicing from DNA sequence using Pangolin. Genome Biol 23:103.
• Dawes R et al. 2023. SpliceVault predicts the precise nature of variant-associated mis-splicing. Nat Genet 55:324.
• Whiffin N et al. 2017. Using high-resolution variant frequencies to empower clinical genome interpretation. Genet Med 19:1151.
• Li MM et al. 2017. Standards and guidelines for the interpretation and reporting of sequence variants in cancer. J Mol Diagn 19:4. (AMP/ASCO/CAP)
• Miller DT et al. 2023. ACMG SF v3.2 list. Genet Med 25:100866.
• Stawiński P, Płoski R. 2024. Genebe.net: implementation and validation of an automatic ACMG variant pathogenicity criteria assignment. Clin Genet 106:119.
• Kopanos C et al. 2019. VarSome: the human genomic variant search engine. Bioinformatics 35:1978.
• Li Q, Wang K. 2017. InterVar: clinical interpretation of genetic variants. Am J Hum Genet 100:267.
• Xiang J et al. 2020. AutoPVS1 -- automated PVS1 decision-tree implementation (verify exact venue/year against the published code/release).
• ClinGen CSpec Registry: https://cspec.genome.network/cspec/ui/svi/all
• ClinGen VCI (Variant Curation Interface): https://curation.clinicalgenome.org/

Related Skills

• clinical-databases/variant-prioritization - Rare-disease pipeline (filters variants; this skill classifies)
• clinical-databases/clinvar-lookup - ClinVar evidence aggregation
• clinical-databases/gnomad-frequencies - BS1/BA1 with Whiffin FAF95
• clinical-databases/myvariant-queries - Aggregated annotation pull
• variant-calling/clinical-interpretation - Clinical reporting workflow