Olto Validation Center

Categorical (Richards Table 5) + Tavtigian points from attributed criteria

Method

Criterion combination is implemented directly from the published rule table and exercised against the worked examples in the source guideline.

Reference

Richards S et al., Genet Med 2015;17(5):405-24 (ACMG/AMP)

Citations

Richards S et al., Genet Med 2015;17:405–424; Tavtigian SV et al., Genet Med 2018;20:1054 / Hum Mutat 2020;41:1734

Sealed fingerprint

ce527ffc577f8b5fbb0838860d21dfea7d30e7b694b02810c87cce38c0ef9c59

Tests

unknown in src/lib/precision/acmg/combine.test.ts

What it does not do

• Criterion ASSIGNMENT is the user’s: the engine combines the criteria it is given, and a wrong criterion in produces a wrong classification out.
• The 2015 guideline has been refined by subsequent ClinGen specifications that are not implemented here.

Physicochemical distance between amino-acid substitutions

Method

Every one of the 190 amino-acid pairs is checked against the published distance matrix.

Reference

Grantham R, Science 1974;185(4154):862-4

Citations

Grantham R, Science 1974;185:862–864

Sealed fingerprint

58c7261f8a85708c892f2b916535694522ebedc5efc0651a012a1edd4dd7c490

Tests

unknown in src/lib/bioinformatics/grantham.test.ts

What it does not do

• A physicochemical distance, not a pathogenicity prediction. A high score is not evidence of effect.

Guruprasad instability index from the dipeptide instability-weight table

Method

The DIWV table and the index formula are checked against the published worked examples.

Reference

Guruprasad K et al., Protein Eng 1990;4(2):155-61

Citations

Guruprasad K et al., Protein Eng 1990;4:155–161

Sealed fingerprint

3d40be70d31cb6c94444632b218e426b2e9c2a8b7b7d094cd6a81ce68c07bd01

Tests

unknown in src/lib/protein-engineering/properties.test.ts

What it does not do

• Predicts in-vitro stability of the primary sequence only; it knows nothing about the folded protein or its expression host.

IWPC square-root weekly dose from demographics + CYP2C9/VKORC1 genotype

Method

The IWPC dose algorithm is reproduced against the coefficients and worked cases in the source publication.

Reference

IWPC, N Engl J Med 2009;360(8):753-64

Citations

International Warfarin Pharmacogenetics Consortium (Klein TE et al.), N Engl J Med 2009;360:753–764

Sealed fingerprint

26ac8b51b575fb41b6191241699fff66324d978407b83dd7f696b3a05b588324

Tests

unknown in src/lib/pgx/warfarin.test.ts

What it does not do

• A starting-dose estimate, not a prescription. Clinical dosing is titrated on INR, and this engine has no access to one.

Calibrated REVEL thresholds → PP3/BP4 strength bands

Method

Threshold banding implemented from the published strength cut-offs. No independent comparison has been run.

Reference

Ioannidis NM et al., Am J Hum Genet 2016;99(4):877-85

Citations

Pejaver V et al., Am J Hum Genet 2022;109:2163–2177 (ClinGen SVI calibration)

Sealed fingerprint

1041d586869f6eb9465bc60625a05124336a52b17862c365e18db29d5bae82ff

Tests

unknown in src/lib/precision/acmg/combine.test.ts

What it does not do

• Bands a score that is computed elsewhere; the score itself is an input, not an output.
• Threshold choices are contested and differ between specifications.

Whiffin maximum-credible-AF + observed-AF → BA1/BS1/PM2 mapping

Method

Maximum credible allele frequency is checked against the worked examples in the source paper.

Reference

Whiffin N et al., Genet Med 2017;19(10):1151-8

Citations

Whiffin N et al., Genet Med 2017;19:1151–1158; Richards S et al., Genet Med 2015 (BA1 5% cutoff)

Sealed fingerprint

11e964f17f4b0897f6d0454b67d798cb87a1b3847dc030c3129a59eff42aebab

Tests

unknown in src/lib/precision/acmg/frequency-exhaustive.test.ts

What it does not do

• Requires prevalence, allelic and genetic heterogeneity and penetrance as inputs. Each is an estimate, and the output inherits every one of them.

Star-allele → activity score → metabolizer phenotype, and CPIC recommendation table

Method

Diplotype to phenotype mapping implemented from the CPIC tables. Exercised against the table, not against an independent implementation.

Reference

CPIC guidelines

Citations

Caudle KE et al., Clin Transl Sci 2020 (CYP2D6 activity-score standardization); CPIC guideline tables (per-gene/-drug)

Sealed fingerprint

ddf9bb409dd2e19fc2ece428f464dce29dbbfc1e577bba86f8495cbe8f38da40

Tests

unknown in src/lib/pgx/star-alleles.test.ts

What it does not do

• Only the genes and alleles present in the implemented tables are recognised; an unlisted allele is reported as unknown rather than guessed.

Isoelectric point, molar extinction (280 nm), GRAVY, aliphatic index, molecular weight

Method

Molecular weight, extinction coefficient and pI are checked against ProtParam outputs for reference sequences.

Reference

Gasteiger E et al., The Proteomics Protocols Handbook, 2005

Citations

Pace CN et al., Protein Sci 1995 (extinction); Kyte J, Doolittle RF, J Mol Biol 1982 (GRAVY); Ikai A, J Biochem 1980 (aliphatic index)

Sealed fingerprint

82a68cb2c124e91f27168baab87a3bfaa638d3b4e047d87811c95c789f56dd7a

Tests

unknown in src/lib/protein-engineering/properties.test.ts

What it does not do

• Computed from the primary sequence: post-translational modification, disulfide state and glycosylation are not modelled.

Wildman–Crippen logP, Ertl TPSA, Lipinski/Veber drug-likeness, and Hückel aromaticity re-perception from a SMILES

Method

SMILES parsing and Crippen logP implemented from the published contribution tables; exercised against a fixed case set.

Reference

Wildman SA, Crippen GM, J Chem Inf Comput Sci 1999;39(5):868-73

Citations

Wildman SA, Crippen GM, J Chem Inf Comput Sci 1999;39:868–873 (atomic-contribution logP; RDKit MolLogP); Ertl P, Rohde B, Selzer P, J Med Chem 2000;43:3714–3717 (TPSA); Lipinski CA et al., Adv Drug Deliv Rev 1997;23:3–25 (Rule of Five); Veber DF et al., J Med Chem 2002;45:2615–2623 (oral bioavailability); RDKit default aromaticity model (Hückel 4N+2), RDKit_Book

Sealed fingerprint

7e5a1a1386c09610c7f69be8c99adfdf1ffd335df83f26022952ac7743178fe9

Tests

unknown in src/lib/molecule/logp.test.ts

What it does not do

• Crippen logP is an atom-contribution estimate and diverges from measured logP for many real compounds.
• Stereochemistry is parsed and preserved but does not affect the computed descriptors.

Deterministic 0–100 target evidence-strength from best study design, corroboration, recency, and contradiction — the authoritative score behind a Verity evidence report (the AI never produces it)

Method

A transparent weighted sum over declared components. Every weight is a published constant in the engine and every input is visible in the output.

Reference

None. This engine has not been compared against an external implementation or dataset.

Citations

OCEBM Levels of Evidence Working Group (Howick J et al.), The Oxford 2011 Levels of Evidence, Oxford Centre for Evidence-Based Medicine; U.S. National Library of Medicine, PubMed/MEDLINE Publication Types (controlled vocabulary)

Sealed fingerprint

2e339c2bff6d0682d52e93413adfca7c9ca5e270586804e7c786a4d0d041326d

Tests

unknown in src/lib/verity/evidence/strength.test.ts

What it does not do

• NOT validated against expert assessment. There is no inter-rater study, no benchmark corpus and no published comparison; the weights are a defensible scheme, not a measured one.
• A score is a summary of what was entered, and cannot detect that a study was entered wrongly.

Deterministic consensus significance + 0–100 evidence-confidence for a gene/variant from ClinVar gold-star review status, submitter concordance, and quote-grounded literature — evidence-gathering, NOT ACMG classification

Method

Weighted combination of ClinVar review status, literature support and concordance. Weights are published constants.

Reference

None. This engine has not been compared against an external implementation or dataset.

Citations

ClinVar review-status / gold-star ratings, U.S. National Library of Medicine (NCBI); Richards S et al., Genet Med 2015;17:405–424 (ACMG/AMP framework; PP5/BP6 deprecation context)

Sealed fingerprint

e54f8bf71aff3b21f67a23f2cb994b0b2c5a701b59d94441cf9ec4fcdeafa899

Tests

unknown in src/lib/verity/variant/strength.test.ts

What it does not do

• Not validated against expert curation. The concordance cap is a design choice, not an empirical finding.

Deterministic 0–100 evidence-landscape strength for a drug × indication from clinical trial phase/breadth/results and quote-grounded literature — describes HOW MUCH evidence exists that the drug was studied, NOT efficacy or a clinical recommendation (literature direction is reported as context, never folded into the score)

Method

Weighted combination of trial phase, breadth and literature support, with every weight published.

Reference

None. This engine has not been compared against an external implementation or dataset.

Citations

U.S. National Library of Medicine, ClinicalTrials.gov (trial registrations, phases, statuses); U.S. FDA / ICH, clinical trial Phase 1–4 framework

Sealed fingerprint

0f1d86fefcec6c5a1937972c3701b9b706a1ed3d1f003fc2f0b782971e1e29f4

Tests

unknown in src/lib/verity/therapeutic/strength.test.ts

What it does not do

• Not validated against clinical outcome. Phase weighting encodes an assumption about evidence value that reasonable people dispute.

Deterministic perfusion-limited physiologically-based pharmacokinetic simulation of a compound through the real circulatory topology (venous → lung → arterial → organs, gut/spleen draining portally through the liver), giving per-organ and plasma concentration-time curves plus non-compartmental PK. Vascular states are blood-referenced and converted to plasma for reporting through an explicit blood:plasma ratio (default 1). Tissue partitioning uses the COMPLETE Poulin & Theil tissue-composition method (phospholipid terms, the fu_p/fu_t binding correction, and the separate vegetable-oil equation for adipose) on the human composition table; it is systematically low for moderate-to-strong bases, which the result flags say. Hepatic elimination can be parameterised three ways, most specific first: saturable Michaelis-Menten on unbound drug (Vmax/Km), unbound-driven linear intrinsic clearance (well-stirred), or a whole-organ clearance. Compound ADME inputs are supplied and are NOT derived here; physiology, the ODE solution and every reported metric are computed. Research use — it predicts exposure under the stated model, it does not establish a dose.

Method

Compartmental ODE integration checked against analytical solutions for the one-compartment cases, and against mass-balance invariants at every step for the rest.

Reference

Analytical one-compartment solutions; conservation of mass

Citations

Rowland M, Peck C, Tucker G. Physiologically-based pharmacokinetic modeling in drug development and regulatory science. Annu Rev Pharmacol Toxicol 2011;51:45; Brown RP et al. Physiological parameter values for physiologically based pharmacokinetic models. Toxicol Ind Health 1997;13(4):407 (organ volumes + blood flows); ICRP Publication 89 (2002), reference anatomical and physiological values; Poulin P, Theil FP. Prediction of pharmacokinetics prior to in vivo studies. J Pharm Sci 2002;91(1):129 (tissue:plasma partitioning); Rowland M, Benet LZ, Graham GG. Clearance concepts in pharmacokinetics. J Pharmacokinet Biopharm 1973;1:123 (well-stirred clearance); Pang KS, Rowland M. Hepatic clearance of drugs. J Pharmacokinet Biopharm 1977;5:625

Sealed fingerprint

e658a6141dc8f5b0a7ffa5473bd4b07689e9c6ec63ce1b50c3eb2a7dba54fa7e

Tests

unknown in src/lib/pbpk/pbpk.test.ts

What it does not do

• Physiological parameters are population defaults. An individual is not a population, and the model has no way to know which.
• Not validated against clinical PK data for any specific drug.

Scales the whole-body simulation to a subject’s covariates (metaboliser activity through the fraction metabolised fm, Child-Pugh hepatic grade, renal function, body weight) and generates a SEEDED virtual population whose 5th/50th/95th-percentile exposure bands are reproducible from the seed alone. Also computes therapeutic-window residence. Covariates scale clearance only; between-subject parameters are currently sampled INDEPENDENTLY, so the bands do not represent covariate correlation.

Method

Applies metabolizer-phenotype and organ-impairment multipliers to the base model. Exercised against the multiplier tables.

Reference

None. This engine has not been compared against an external implementation or dataset.

Citations

Caudle KE et al. Standardizing CYP2D6 genotype to phenotype translation. Clin Transl Sci 2020;13:116 (activity-score framework); Verbeeck RK. Pharmacokinetics and dosage adjustment in patients with hepatic dysfunction. Eur J Clin Pharmacol 2008;64:1147 (Child-Pugh); Rowland M, Tozer TN. Clinical Pharmacokinetics and Pharmacodynamics (renal clearance proportional to GFR); Jamei M et al. The Simcyp population-based ADME simulator. Clin Pharmacokinet 2009;48:307 (virtual-population framework)

Sealed fingerprint

18fbad34fad21df84e542ec4f71d8e5127525f4cf11b6c563fc2295e4b93a300

Tests

unknown in src/lib/pbpk/precision.test.ts

What it does not do

• Multipliers are literature-derived point estimates with real between-study variation that is not propagated.
• Not validated against measured exposure in impaired or variant populations.

Simulates a perpetrator compound’s own pharmacokinetics, then re-simulates the victim under the resulting time-varying hepatic clearance, reporting AUC and Cmax ratios and an FDA-threshold classification. The same clearance-scaling contract carries a star-allele diplotype through the CPIC activity score to a whole-body exposure prediction. The interaction is applied through the fraction of the victim’s hepatic clearance the affected enzyme carries (fm), so the AUC ratio is bounded by 1/(1−fm) as inhibition becomes complete rather than rising without limit. fm defaults to 1 — the whole hepatic clearance responding — which is an upper bound unless a victim-specific fm is supplied.

Method

Interaction modelling over the base PBPK engine. Exercised for internal consistency only.

Reference

None. This engine has not been compared against an external implementation or dataset.

Citations

FDA Guidance for Industry. Clinical Drug Interaction Studies — Cytochrome P450 Enzyme- and Transporter-Mediated Drug Interactions (2020); Fahmi OA et al. Comparison of different algorithms for predicting clinical drug-drug interactions. Drug Metab Dispos 2009;37:1658; Caudle KE et al., Clin Transl Sci 2020 (activity-score standardization); CPIC gene-drug guidelines

Sealed fingerprint

ca31e6365c03595ab0d37c7588038ce856e328f725bade8b00cb09ab5b3877ab

Tests

unknown in src/lib/pbpk/ddi.test.ts

What it does not do

• EXPERIMENTAL. Not represented as validated, and not for clinical or regulated use.
• Mechanism coverage is partial: induction and time-dependent inhibition are simplified.