Phage and yeast display

Display links a protein to the genetic material encoding it, so that selecting the protein recovers the gene. Phage display puts the protein on a filamentous phage coat; yeast display puts it on the yeast cell wall. Both let you search libraries far larger than you could ever express and test individually.

What each platform is good at

Phage display handles very large libraries, is inexpensive per round, and works well for fragments such as scFvs, Fabs and VHHs. Selection is done by panning against an immobilized or biotinylated target, with washes providing the pressure.

Yeast display handles smaller libraries but gives a quantitative, per-cell readout, because the displayed protein and the bound target can both be labeled and measured by flow cytometry. That makes it excellent for affinity maturation, for normalizing to expression level, and for sorting on a defined threshold rather than on a wash.

They combine well: phage to reduce a large library to a manageable population, yeast to refine it with quantitative sorting.

Selection pressure is the experiment

What you get out is decided by what you selected for.

• Target concentration. Lowering it in successive rounds selects for higher affinity.
• Wash stringency and off-rate selection. Long washes, or competition with excess unlabeled target, select for slow dissociation.
• Valency. Multivalent display and multivalent target both create avidity, which hides weak affinity. Monovalent display is what you want when affinity is the goal.
• Counter-selection. Depleting on a related protein, on the tag, on streptavidin or on the plastic removes the binders you do not want. Anti-streptavidin and anti-tag binders are a common and avoidable outcome.
• Negative selection on cells removes polyspecific binders early, which saves a great deal of downstream disappointment.

The biases to expect

Every round enriches for expression and stability as well as for binding, because a clone that displays poorly cannot be selected. That is partly useful, since it favors well-behaved molecules, and partly a distortion, since it can lose good binders that display badly.

Phage panning also enriches for whatever sticks: plastic, blocking agent, the capture reagent. Rotating blocking agents and capture chemistry between rounds is a cheap and effective countermeasure.

Reading the output

Sequencing the population after each round tells you far more than picking clones does. Enrichment trajectories across rounds separate real binders from lucky survivors, and clustering by CDR3 shows how much diversity remains. Over-selecting produces a population dominated by one or two clones, which looks like success and leaves nothing to engineer.

Confirm hits as soluble protein. Behavior on a phage or a yeast surface does not always survive expression and purification, and the confirmation step is where the real panel begins.