Question 1

Why are pharmacogenes difficult to genotype accurately?

Accepted Answer

Several key genes sit near highly similar pseudogenes, carry hybrid arrangements, and vary in copy number, which short-read sequencing resolves poorly. Targeted assays or long-read data handle these regions far better. State which alleles the method used could and could not detect, since undetected alleles are silently assigned as normal function.

Question 2

How is genotype converted to a predicted phenotype?

Accepted Answer

Each allele receives an activity value, the two values combine into a score, and the score maps onto a metaboliser category. Because scoring schemes and category boundaries differ between sources, name the scheme applied and keep it consistent across the study. Reporting the underlying diplotype lets others reinterpret under a different scheme.

Question 3

Why does ancestral diversity matter so much here?

Accepted Answer

Clinically important alleles differ in frequency between populations, and some are common in groups poorly represented in reference studies. A panel designed around one population misclassifies others as normal function. Report the ancestral composition of the cohort and whether the assay covers alleles relevant to the groups included.

Question 4

What confounding threatens observational outcome studies?

Accepted Answer

Prescribing choice reflects patient characteristics, so the sickest patients often receive particular drugs, creating confounding by indication. Adherence, interacting medicines, kidney and liver function, and comorbidity all influence response independently. Pre-specify covariates, consider active comparator designs, and treat unadjusted associations as hypothesis-generating rather than conclusive.

Question 5

What distinguishes an implementation study?

Accepted Answer

It asks whether delivering the result changes prescribing and outcomes, not whether the association exists. Endpoints include time to result availability, proportion of prescriptions following guidance, and clinical events. Randomising at clinic or cluster level is common, and process measures matter as much as outcomes because most failures occur in delivery.

Question 6

How large must a cohort be?

Accepted Answer

Size follows the frequency of the variant carrier group and the expected effect. Rare alleles with strong effects need enrichment strategies or consortium collaboration rather than single-centre recruitment. Power calculations should use realistic effect sizes from prior studies, since optimistic assumptions produce cohorts unable to answer the question posed.

Question 7

What governance applies to this work?

Accepted Answer

Ethics approval, consent covering genetic analysis and secondary use, controlled data access and a plan for handling findings unrelated to the study question must be settled before recruitment. Because results may remain clinically relevant for a lifetime, retention and future return of results deserve explicit discussion in the protocol.

Biotechnology

Biotechnology

Research Frontiers in Precision Medicine and Pharmacogenomics