AI in Genomics Quiz

Answers and explanations appear as you take the quiz.

1. Roughly how many base pairs are in one haploid human reference genome?
   • About 3 billion
   • About 3 thousand
   • About 3 million
   • About 3 trillion
2. What does DeepMind's AlphaMissense predict?
   • The price of gene therapies
   • Whether missense mutations are likely benign or pathogenic
   • A patient's blood type
   • The shape of red blood cells
3. Why is DNA often treated like a 'language' by AI models?
   • Because it is written in English
   • Because genes can talk
   • Because it is a long sequence of letters (A, C, G, T) with learnable patterns
   • Because it uses punctuation marks
4. What is a polygenic risk score?
   • A test for a single broken gene
   • A measure of how many genes you have
   • A score for DNA sequencing speed
   • An estimate of disease risk combining many small genetic effects
5. In AlphaMissense, how does evolutionary conservation function in the model?
   • It is one signal combined with structural context to help predict missense effects
   • It proves every change at a conserved position is harmful
   • It alone determines a clinical diagnosis
   • It measures a patient’s disease severity
6. What is the goal of AI designing CRISPR guide RNAs?
   • To make DNA glow
   • To maximize on-target edits and minimize off-target effects
   • To slow down cell division
   • To translate DNA into protein
7. What information did AlphaMissense use in adapting an AlphaFold-based model for missense-effect prediction?
   • Only a patient’s symptoms
   • Only CT images
   • Human and primate variant population-frequency data
   • A list of medical billing codes
8. How should a predicted pathogenicity score be used in clinical variant interpretation?
   • As a stand-alone diagnosis
   • As proof that a person will develop disease
   • As a replacement for genetic counseling
   • As one computational evidence source alongside clinical and functional evidence