Sztuczna inteligencja w quizie dotyczącym farmakogenomiki

Odpowiedzi i wyjaśnienia pojawiają się w miarę rozwiązywania quizu.

1. A CPIC guideline exists for a gene-drug pair. According to the guide, what question does it answer?
   • Whether a clinician should order the genetic test
   • How to adjust prescribing once a test result is already available
   • Which lab should run the test
   • How much the test should cost
2. A patient's genotype shows normal CYP2D6 function, but they take paroxetine. What should the pharmacist consider?
   • The patient may behave like a poor metabolizer because of phenoconversion
   • The genotype result must be a lab error
   • Paroxetine is removed from the body entirely through CYP2C19
   • The patient is automatically an ultrarapid metabolizer
3. Why does the guide warn against giving codeine to a CYP2D6 ultrarapid metabolizer?
   • Codeine does nothing at all in these patients
   • They turn codeine into morphine quickly, which risks toxicity
   • Codeine permanently blocks the CYP2D6 gene
   • The HLA-B gene forbids it
4. How is a CYP2D6 activity score calculated, as described in the guide?
   • By counting the patient's current medications
   • By measuring morphine in a blood sample
   • By adding the activity values assigned to the two star alleles
   • By averaging results from three separate labs
5. What makes CYP2D6 especially hard to call from short-read sequencing?
   • It sits on the Y chromosome only
   • It has no known variants
   • Deletions, extra copies and hybrids with the similar CYP2D7 pseudogene
   • It is only expressed in the brain
6. Which misconception about pharmacogenomic panels does the guide correct?
   • That they can show which antidepressant will work for a person
   • That they involve DNA
   • That CYP2C19 affects clopidogrel
   • That HLA-B variants relate to drug reactions
7. Why might a targeted genotyping panel wrongly report a normal-function *1 allele?
   • It sequences every base of the gene twice
   • It only tests preselected variants, so an untested rare variant goes unseen
   • It always reports *1 for women
   • It relies on the patient's family history
8. The guide says decision-support systems often store a coded phenotype rather than raw variants. What maintenance risk does this create?
   • Phenotypes take up too much storage
   • Stored phenotypes can become outdated when CPIC changes its translation tables
   • Coded phenotypes cannot trigger alerts
   • Pharmacists are not allowed to read phenotypes