Deep mutational scanning has become a core method in therapeutic antibody engineering because it answers, in one position-resolved experiment, questions that used to take many. A single saturation scan can rank affinity-improving mutations, expose developability trade-offs, map the functional epitope, and predict escape — depending on what you scan and how you select. This article covers the four ways DMS supports an antibody program and how the campaigns are built.
One method, four antibody questions
The power of DMS for antibodies is that the same experimental machinery — a saturation library, a functional selection, and an NGS readout — can be pointed at different questions by changing what is scanned and what is selected. The general method is covered in deep mutational scanning: mapping protein fitness landscapes; here is how it specializes for therapeutic antibodies.
1. CDR scanning for affinity
Saturate every CDR position, screen the library under binding selection, and read out a complete single-mutant map of which substitutions improve affinity at which positions. That map turns affinity maturation from a random walk into a ranked plan: combine the top beneficial mutations into a focused second-round library instead of guessing. This is the maturation workflow detailed in DMS for antibody affinity maturation.
2. Developability co-optimization
Affinity is only useful if the molecule remains a drug. Because surface expression on a display platform correlates with folding stability, scanning for expression in parallel with binding flags the positions where a tighter-binding mutation would cost stability, yield, or aggregation resistance. The output is a trade-off map: which affinity gains are free and which come with a developability tax. That visibility matters because developability liabilities are a leading cause of late-stage antibody attrition, as we cover in five developability red flags that kill mAb programs.
3. Functional epitope mapping
Point the scan at the antigen instead of the antibody. Saturate the target surface, select for binding to the therapeutic antibody, and the loss-of-binding mutations trace the functional epitope at single-residue resolution. This is more precise than competition binning: it tells you which residues the antibody actually depends on, which supports epitope-binning decisions, differentiation from competitor antibodies, and novel-composition intellectual-property claims.
4. Escape and resistance mapping
For rapidly evolving targets, the most valuable question is which antigen mutations let the target escape the antibody. Saturate the antigen, select against the therapeutic antibody, and the variants that retain function while losing antibody binding are the escape map. This approach was applied at scale to SARS-CoV-2 receptor binding domain antibodies and generalizes to influenza, other viral targets, and any program where resistance must be anticipated. The escape map feeds directly into designing escape-resistant or broadly neutralizing leads.
How the campaigns are built
Whichever question is being asked, the campaign shares a structure:
- Library design. Saturation mutagenesis across the relevant region — CDRs for antibody optimization, the antigen surface for epitope and escape work — with diversity matched to selection throughput so every variant is sampled many times.
- Functional selection. The library is displayed on yeast or mammalian cells and sorted for the phenotype of interest: binding, expression, or both. Stringency tunes which substitutions register as gain or loss.
- NGS readout. Pre- and post-selection pools are sequenced, with UMI tagging and replicates to control sampling and PCR noise.
- Fitness scoring. Log-enrichment ratios per variant, normalized to wild type, produce the position-by-substitution matrix that downstream design consumes.
The platform choice follows the antibody format and the epitope. Small formats and soluble antigens scan well on yeast display; glycan-dependent epitopes and full-length-IgG context move to mammalian display.
Where it fits in a program
DMS is the evidence layer beneath antibody engineering: it tells you where to mutate, what it will cost, what the antibody binds, and how the target can escape. Paired with the wet-lab optimization on antibody and nanobody engineering, it turns each of those questions from a guess into a measurement.
To scope a deep mutational scanning campaign for an antibody program, see our deep mutational scanning services and antibody engineering services, or start a project.
Related Ranomics services
- Deep mutational scanning: Library design, selection, NGS, and fitness scoring as a CRO service.
- Antibody and nanobody engineering: The wet-lab optimization that DMS data drives.
- Affinity maturation: Turn a CDR scan into a ranked, combinatorial maturation campaign.
Frequently asked questions
What can deep mutational scanning do for a therapeutic antibody?
Four things from the same kind of experiment: rank every affinity-improving CDR substitution, co-score developability liabilities, map the functional epitope by scanning the antigen, and identify the antigen mutations that let a target escape the antibody. It replaces sequential single-mutant assays with one position-resolved fitness map.
How is DMS different from scanning just for affinity?
Affinity maturation is one use of DMS. For a therapeutic program, the same scanning machinery also answers developability, epitope, and escape questions. Scanning the antibody optimizes the binder; scanning the antigen characterizes the interaction and its vulnerabilities.
Can DMS map antibody escape mutations?
Yes. By saturating the antigen and selecting against the therapeutic antibody, DMS reveals which target mutations abolish binding — the escape map. This was used extensively for SARS-CoV-2 receptor binding domain antibodies and applies to any rapidly evolving target where resistance is a concern.
Do you run this as a service?
Yes. We run end-to-end deep mutational scanning campaigns: library design, functional selection on yeast or mammalian display, NGS, and a normalized fitness matrix with ranked variant lists, scoped to the antibody-engineering question you are asking.