AI in Drug Discovery Quiz

Answers and explanations appear as you take the quiz.

1. What did DeepMind's AlphaFold accomplish that aids drug discovery?
   • It predicted the 3D folded structures of over 200 million proteins
   • It synthesized new drugs in a lab
   • It ran human clinical trials automatically
   • It replaced all chemists
2. How are molecules commonly represented for machine learning models?
   • Only as photographs
   • As graphs of atoms and bonds or as SMILES text strings
   • As audio waveforms
   • As spreadsheets of prices
3. What is 'virtual screening' in AI drug discovery?
   • Watching a movie about drugs
   • Screening patients for a disease
   • Computationally ranking millions of compounds to find promising ones without synthesizing them all
   • A type of video call with regulators
4. What can an ADME or toxicity prediction contribute during early candidate assessment?
   • It proves a molecule is safe in humans
   • It eliminates the need for animal or laboratory studies
   • It guarantees the candidate will pass clinical trials
   • It can help prioritize candidates for experimental follow-up, but does not establish safety or replace testing
5. Which generative AI techniques are used to design novel molecules?
   • Variational autoencoders and diffusion models
   • Spreadsheet macros
   • Barcode scanners
   • Email filters
6. What must generally occur before a new drug candidate can begin clinical research in people in the United States?
   • A model score above 0.8
   • Preclinical safety information and an FDA-reviewed Investigational New Drug application
   • A successful virtual screen alone
   • Commercial availability in a pharmacy
7. What does AlphaFold 3 add beyond predicting protein structures?
   • It proves a predicted drug will succeed in patients
   • It replaces all clinical trials
   • It can predict structures and interactions involving proteins, nucleic acids, and ligands
   • It estimates only the retail price of a medicine
8. What does a virtual-screening rank mean for a candidate compound?
   • The compound is already an approved drug
   • The compound is guaranteed safe in humans
   • The model has observed the compound in a clinical trial
   • It is a computational priority for follow-up, not proof of efficacy