MCHM3001 · From Molecules to Therapeutics
Drug Discovery: History & the Modern Pipeline
The opening lecture of University of Sydney MCHM3001 frames the whole unit: what a drug is, how discovery moved from natural products and serendipity to a rational target-to-market pipeline, and the cautionary lessons (thalidomide) that shaped modern regulation. It also introduces the decision that recurs all semester — structure-based versus ligand-based design — and the numbers of the druggable genome. This material sets the vocabulary the rest of the pipeline is written in, and shows up in Test 1 and the comprehensive final.
What this chapter covers
- 01Definition of a drug; the Doctrine of Signatures and natural products (opium alkaloids, aspirin lineage: salicin → salicylic acid → acetylsalicylic acid)
- 02Serendipity in discovery (ephedrine, LSD, chlordiazepoxide, artificial sweeteners) and the thalidomide teratogen lesson (in-vivo racemisation of the (S)-enantiomer)
- 03Classical (lead → analogues → phenotypic screen) versus modern (genome → target → HTS → optimisation → trials) discovery
- 04The modern pipeline stages: target identification and validation → hit discovery → hit-to-lead → lead optimisation → preclinical → Phase I/II/III → approval
- 05The druggable genome (Hopkins & Groom): ~3,000 druggable proteins ∩ ~3,000–10,000 disease genes ≈ 600–1,500 targets; target-class split (enzymes, GPCRs, ion channels)
- 06SBDD versus LBDD decision tree (3D protein structure known or not)
- 07Mechanisms of drug action: agonist, partial agonist, inverse agonist, competitive and irreversible antagonist; allosteric PAM/NAM/SAM
- 08R&D economics teaser: cost per new drug and Eroom's law
Choosing a discovery path and sizing the target space
- +1(a) A solved 3D structure means you can see the binding site, so use Structure-Based Drug Design (SBDD) — docking, virtual screening and fragment methods. Ligand-Based Drug Design (pharmacophores, SAR) is the fallback used only when no 3D structure is available.
- +1(b) Order the pipeline: target identification and validation → hit discovery (library + screening) → hit-to-lead → lead optimisation → preclinical development → clinical Phase I → Phase II → Phase III → regulatory approval and market.
- +1(c) State the two sets: the druggable genome is ~3,000 proteins that can bind a drug-like small molecule; disease-modifying genes number ~3,000–10,000.
- +1(c cont.) A viable target must be in BOTH sets, so take the intersection, not either whole set — this is the key conceptual move.
- +1(c cont.) The intersection Hopkins & Groom estimate is roughly 600–1,500 proteins — the realistic universe of drug targets, far smaller than either the druggable set or the disease set alone.
Key terms
- Drug
- Any synthetic, semisynthetic or natural chemical substance, excluding food, that interacts with a biological system to affect its physiological function(s).
- Druggable genome
- The ~3,000 human proteins able to bind a drug-like small molecule with useful affinity (Hopkins & Groom, 2002); only a fraction of the ~20,000-protein proteome.
- Drug target
- A protein that is both druggable and disease-modifying; the intersection of the druggable and disease sets is estimated at ~600–1,500 proteins.
- SBDD vs LBDD
- Structure-Based Drug Design uses a known 3D protein structure (docking, virtual screening, fragments); Ligand-Based Drug Design is used when no structure is available (pharmacophores, SAR, rational design).
- Agonist / antagonist
- An agonist binds and produces a response (full, partial, or inverse for constitutive activity); an antagonist binds without intrinsic effect and blocks the response (competitive/reversible or irreversible/covalent).
- Thalidomide lesson
- Sold as a racemate; the (S)-enantiomer is a potent teratogen and the drug racemises in vivo, so single-enantiomer dosing does not help — a landmark case that tightened drug regulation.
Drug Discovery: History & the Modern Pipeline FAQ
What is the difference between the druggable genome and the set of drug targets?
The druggable genome is every protein that can physically bind a drug-like small molecule with useful affinity — about 3,000 human proteins. A drug target must additionally be disease-modifying, so the real target universe is the intersection of the druggable set (~3,000) with the disease-gene set (~3,000–10,000), estimated at roughly 600–1,500 proteins. A protein can be very druggable yet useless as a target if hitting it does not change disease.
When do you use structure-based versus ligand-based design?
If you have a three-dimensional structure of the target (from X-ray crystallography, cryo-EM or NMR), use structure-based design — you can dock candidates and run virtual screening against the actual binding site. If no structure is available, fall back to ligand-based design: build a pharmacophore from known active molecules and use structure-activity relationships. This choice recurs across the unit.
Why is thalidomide such an important example?
It shows why chirality and metabolism matter. Thalidomide was sold as a racemate; the (S)-enantiomer is a potent teratogen, and crucially the molecule racemises in the body, so dosing only the 'safe' (R)-enantiomer would not have prevented harm. The disaster (10,000–20,000 affected births) drove the post-1962 tightening of drug regulation and preclinical testing.
Can AI help me with the drug-discovery pipeline in MCHM3001?
Yes. Sia can recite and explain the pipeline stages in order, drill you on the SBDD-versus-LBDD decision, or quiz you on binding modes (agonist, partial agonist, inverse agonist, competitive and irreversible antagonist, and allosteric PAM/NAM/SAM). It explains the reasoning and checks your answers so you learn the framework; it will not sit an assessment for you, and University of Sydney academic-integrity rules apply.
Exam move
Learn the modern pipeline as an ordered spine you can recite — target ID and validation, hit discovery, hit-to-lead, lead optimisation, preclinical, Phase I/II/III, approval — because every later chapter slots onto it. Pair each historical example with the principle it teaches (aspirin = lead development from a natural product; thalidomide = chirality plus in-vivo racemisation; the sweeteners and LSD = serendipity). Memorise the druggable-genome triangle (druggable ~3,000 ∩ disease ~3,000–10,000 ≈ 600–1,500 targets) and the SBDD-versus-LBDD decision, both of which make easy short-answer marks in Test 1 and the final. When a binding-mode distinction blurs, ask Sia to contrast, say, a partial agonist with a competitive antagonist using a worked receptor example.
Working through Drug Discovery: History & the Modern Pipeline in MCHM3001? Sia is AskSia’s AI Chemistry tutor — ask any MCHM3001 Drug Discovery: History & the Modern Pipeline question and get a clear, step-by-step explanation grounded in how MCHM3001 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
