recordRiskGene

a gene whose clinical consequence is allele CARRIAGE, not enzyme activity

Information

CPIC assigns these genes by ALLELE STATUS rather than by a metaboliser phenotype (its lookupmethod field), and they do not belong in the Gene/Effect machinery: carrying HLA-B*57:01 does not scale abacavir's clearance or its EC50 by any factor. It means the drug must not be given.

Why carriage is a Boolean and not a risk factor

A scale factor asserts a dose-response relationship — twice the parameter, twice the effect — and for an idiosyncratic immune reaction there is none: there is no dose at which a *57:01 carrier is safe, and no multiplier on a PK or PD parameter reproduces that. Modelling carriage as a factor would put a number into a differential equation that the underlying biology does not support. So isCarrier returns a Boolean, CarriageComponent emits it as a flag and warns, and nothing here multiplies anything.

oddsRatio exists so a published association can travel with the record for reporting and for cohort-level work. It is not read by any equation. If you need it in a calculation, that calculation is an epidemiological one about a population, not a pharmacokinetic one about this subject.

Genes with GRADED deficiency (G6PD's class I-IV, for instance) fit neither this record nor Gene cleanly: the severity classes are ordered but they are not metaboliser activity. They are reported by the generator as unmapped rather than forced into either.

Parameters

TypeNameDefaultDescription
Stringsymbol""HGNC symbol, e.g. "HLA-B"
StringriskAllelefill("", 0)alleles that carry the risk, e.g. {"*57:01"}
Stringoutcome""what carriage risks, e.g. "hypersensitivity reaction"
RealoddsRatio1.0reported odds/hazard ratio, 1 when not quantified
StringnegativeLabel"negative"the explicit not-a-carrier value