Question 1

How is the human scaffold chosen?

Accepted Answer

Three routes are common: the closest matching human sequence, a consensus sequence for the family, and the nearest inherited sequence before any maturation. The inherited route gives the lowest recognition risk but often the largest binding loss. Try more than one scaffold in parallel rather than committing to a single choice.

Question 2

Why does grafting the contact loops reduce binding?

Accepted Answer

Positions in the scaffold immediately beneath the loops set their angle and packing, and these differ between the original and human sequence. The loops therefore adopt slightly different shapes on the new scaffold. Identifying which of those supporting positions must revert to the original residue is the central design task.

Question 3

How much binding loss is acceptable?

Accepted Answer

It depends on the starting affinity and the target concentration, not on a fixed rule. A modest loss from a very tight parent may be irrelevant, while the same loss from a moderate binder is fatal. Measure the parent and all variants by one method and report the values rather than fold-changes alone.

Question 4

How is immune recognition assessed before the clinic?

Accepted Answer

Combine prediction of peptide binding to human presentation molecules with assays using cells from donors covering common tissue types. Both give useful signals and neither predicts patient response reliably, so report them as risk indicators rather than as clearance, and prefer designs that reduce predicted sites without harming binding.

Question 5

What developability properties must be screened?

Accepted Answer

Expression level, thermal stability, tendency to clump at the concentrations needed for injection, solution viscosity, and sequence positions liable to chemical change during storage. Screen these alongside binding from the first variant round, since discovering a formulation problem after selecting a lead wastes a great deal of work.

Question 6

How should human sequence content be reported?

Accepted Answer

Give the percentage identity to the closest human sequence for the variable regions, state how many non-human residues were retained outside the contact loops and why, and specify the reference set used. Describing a molecule as humanised without these numbers conveys very little to a reader.

Question 7

When is a fully human binder the better route?

Accepted Answer

Where display libraries built on human sequence or animals carrying human antibody genes can supply a binder against the target directly, that route avoids the conversion problem entirely. Conversion remains valuable when an existing non-human binder has properties that have proved difficult to obtain by other means.

Biotechnology

Biotechnology

Research Frontiers in Therapeutic Antibody Humanization