Instructor Notes — Chapter 35

Peptide Discovery — From Venom to Artificial Intelligence Part 6 · intermediate · prerequisites: Chapters 33, 2, 5

What this chapter is actually for

On the surface it is a survey: four discovery routes, some charismatic animals, some Nobel Prizes. Taught that way it is pleasant and forgettable.

The chapter's real job is to install one idea: that improving a stage of a process is not the same as improving the process. Everything before §35.10 is evidence-gathering for that point, and §35.10 is where the chapter cashes it in. If your students leave able to recite the four routes but unable to explain why faster discovery has not produced proportionally more drugs, the session did not land.

The secondary job is to give the book's most delightful material without letting delight substitute for rigor. Venoms, frog skin, cone snails, a lizard that beat a decade of medicinal chemistry — this is the chapter students will actually enjoy, and that enjoyment is the point of vulnerability. The whole dossier coda exists because charming origin stories are the most effective substitute for evidence in this entire field.

Let the delight show. Then close the trap.

Timing and shape

A single 90-minute session works if you accept that §35.6 and §35.7 get compressed. A two-session split is better:

• Session 1 — nature. §35.1–§35.5, anchored on the Gila monster. Case study 35.1.
• Session 2 — technology. §35.6–§35.10, anchored on AlphaFold. Case study 35.2, then §35.10 as the synthesis.

If you have only one session, cut §35.6 and §35.7 to ten minutes combined and protect §35.10. The display and cryo-EM material is genuinely interesting and it is not what the chapter is about.

The three things students most often get wrong

1. They attribute semaglutide's albumin trick to the lizard.

This is the single most common error and it appears in exam answers every time. Exendin-4 solved DPP-4 cleavage. It did nothing about renal clearance. The fatty acid chain that extends half-life from hours to days has no natural precedent and came from the laboratory. Quiz item 5 is built to catch this; if more than a third of the room misses it, spend five minutes on the two-clearance- mechanism picture before moving on.

2. They convert "nature knows best" into a general principle.

The exendin-4 story is a very good story for the naturalist position and the chapter concedes it fully. Then the same story continues past the point where it is usually stopped. Students who stop early will produce essays arguing that pharmaceutical companies should redirect resources toward bioprospecting. That is not an absurd position, but it is not what this evidence supports, and the exercise is to notice that the evidence they are citing contains its own rebuttal.

3. They over-correct on AlphaFold.

Having learned that hype exists, a certain kind of good student will decide AlphaFold is overrated. It is not, and the chapter is emphatic about this — the technical claim is rated ✅ on the strength of a blind assessment better controlled than many clinical trials. Do not let the room leave in a posture of general skepticism about technology. The skill being taught is scoping a claim, not disbelieving claims. A student who cannot articulate why the CASP14 result is excellent has not learned the lesson; they have learned a mannerism.

Board work worth doing live

The two-clearance-mechanism diagram. Draw native GLP-1 with two arrows pointing at it: DPP-4 cleavage and renal filtration. Then show exendin-4 blocking one, and semaglutide blocking both. Ten seconds of chalk saves twenty minutes of confusion.

The pipeline funnel from §35.10. Draw it before you say anything about AI. Then ask the room to point at where all the technology in the chapter acts, and separately at where the compounds are lost. Let the gap between the two fingers be the argument. This works far better as a live demonstration than as a read diagram.

The captopril / orforglipron parallel. Two three-step arrows stacked. Students who have read Chapter 33 recognize the second arc immediately, and the recognition does more teaching than the explanation does.

Assessment notes

• Quiz item 20 (which is not an AlphaFold limitation) is reversed on purpose and is the best single discriminator in the set.
• Exercises marked † are suitable as written assignments. C, K, W, AE, and AH are the strongest for graded work; AE in particular (apply the bottleneck rule outside medicine) reliably separates students who have internalized the idea from those who have memorized the sentence.
• Case study 35.2 works as a take-home applying Chapter 5's method to a non-drug claim. Require the ✅ to be stated before the ⚠️ — this is the structural fix for over-correction.
• The dossier exercise (Field 12 specification) is the highest-value practical assignment in Part VI and is worth grading properly. Insist on both thresholds. A specification with only an upgrade condition is the tell that the student is hoping rather than evaluating.

Sensitivities

Nothing in this chapter is clinically fraught, with two small exceptions.

Ziconotide is discussed in the context of severe refractory chronic pain, and some students will have personal experience in that territory. Handle the "implanted pump" material as what it is — a description of the cost of a delivery constraint, not a judgment about anyone's treatment. The safety callout ends by returning decisions to clinicians; keep that framing in discussion.

GLP-1 nausea comes up in §35.5 and in exercise T. Students on these drugs are in most rooms now. The on-target framing is genuinely reassuring rather than dismissive — it means the effect is understood, not that it is imaginary — but it needs to be said in a way that does not sound like the side effect is being minimized.