Home › Services › Thermostability Engineering

Can you select for a heat-stable antibody instead of predicting one?

Yes. Rather than computing which mutations might raise melting temperature and hoping the prediction survives contact with the bench, Abwiz Bio makes thermostability a condition of survival. We transiently heat and cool the phage library between panning rounds: clones that unfold irreversibly are lost, thermoresistant clones are recovered and amplified, and stringency is raised at each successive stage. Because panning against your antigen continues throughout, stability is gained without trading away binding — clones that lose the parent specificity do not survive the selection. The method is established and peer-reviewed (2016), and the same filter runs as standard inside every STEM™ engineering campaign.

Why thermostability decides a program late

Thermal stability is rarely the reason a clone is chosen, and frequently the reason one is abandoned. An antibody that unfolds at low temperature aggregates during purification, loses titer in the bioreactor, degrades on the shelf, and fails freeze–thaw. In diagnostics it dictates whether a reagent can be shipped without a cold chain or lyophilized into a kit. None of these show up in the ELISA that selected the lead.

The conventional response is computational: scan the sequence, predict destabilizing residues, mutate them, and measure the result. That works sometimes, but each predicted mutation is a fresh risk to the paratope, and the cycle of design, express, and measure is slow enough that only a handful of variants ever get tested.

How selection-based thermostability works

Phage display lets stability be tested on millions of clones at once instead of a dozen.

• Heat stress between rounds. The library is transiently heated and then cooled before selection. Clones whose displayed Fab or VHH cannot refold are eliminated from the pool; those that survive are, by definition, the ones that recovered.
• Escalating stringency. Temperature and exposure are increased stage by stage, so the surviving population is progressively enriched for the most thermoresistant members rather than merely for the merely-adequate.
• Function held constant. Antigen panning runs alongside the heat challenge, so a clone must be both stable and still binding to make it through. Stability that costs affinity does not survive the selection.
• Confirmation after IgG conversion. Because stability was selected on the folded, displayed domain, the property carries into the full-length molecule; we confirm it directly on purified IgG rather than assuming it.

When the starting lead needs more sequence diversity than it natively contains, we combine the heat challenge with STEM™ library expansion, where CDR positions are designed from human antibody amino-acid usage with a local AlphaFold instance plus germline and structural bioinformatics. Polyreactivity filters run in parallel, so the output is stable, specific and developable at once.

What this connects to

Thermostability is one axis of a larger biophysical picture, and it is usually engineered alongside others:

• Developability optimization — aggregation, high-concentration viscosity, sequence liabilities, PTM susceptibility and polyreactivity, engineered on the same library.
• Expression and titer rescue — poorly folding variable domains are frequently both unstable and low-expressing, and one selection can address both.
• VHH and single-domain antibodies — single domains are intrinsically robust, and heat selection sharpens that advantage further.

Our own published work on rabbit antibody thermal behaviour is summarized in the thermal-stability study, and the solubility characteristics of engineered single domains in the single-domain solubility study.

You receive

• Thermostability data for selected clones against the parent
• Developability profile covering the biophysical properties measured during selection
• Binding data confirming retained function (ELISA; BLI kinetics optional after IgG purification)
• Full VH/VL or VHH sequences and recombinant IgG
• No downstream royalties, stage-gated approval before each next step

Frequently asked questions

How do you select for thermostability with phage display?

The phage library is transiently heated and then cooled between panning rounds. Clones that unfold irreversibly are lost from the pool, while clones that refold survive and are amplified. Stringency is raised at each successive stage, so the population is progressively enriched for the most thermoresistant members. Antigen panning continues throughout, so surviving clones must remain functional as well as stable.

Will making the antibody more stable reduce its affinity?

It is designed not to, because binding is a survival condition rather than something measured afterwards. Clones are panned against your antigen in the same rounds in which they are heat-challenged, so a variant that gains stability by damaging the paratope is removed by the selection itself. Where a clone does trade some affinity for stability, that shows up in the data delivered against the parent rather than being assumed away.

Is this different from computational stability prediction?

Yes. Predictive methods propose a small number of stabilizing mutations that must then be built and measured one at a time. Selection tests stability empirically across the entire library simultaneously, so the answer comes from the bench rather than from a model, and rare stabilizing combinations that no predictor would have proposed can still be recovered.

Does the stability measured on phage carry over to full-length IgG?

It generally does, because the property is selected on the folded, displayed antibody domain rather than inferred from sequence. We confirm it directly on purified IgG after conversion, so the delivered data describes the molecule you will actually use rather than the phage.

Is the method published?

Yes, the approach is established in the peer-reviewed literature (2016) and has been applied across Abwiz Bio campaigns since. It also runs as a standard filter inside every STEM engineering project, not only in dedicated thermostability work.

Can you engineer thermostability into an antibody I already have?

Yes. Send the VH/VL or VHH amino-acid sequence; no immunization or new discovery campaign is needed. Contact info@abwizbio.com with your sequence and the stability target your application requires for a scoped quote.