Pin-Kuang Lai

speaker

67 appearances 1 recordings 1 series first heard Jul 2026 last heard 23 Jul

Pin-Kuang Lai’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

Talking Techniques · AI & Antibodies miniseries | The series concludes: our review with Guest Advisor Pin-Kuang Lai · 23 Jul 2026 podcast

The community increasingly recognizes that identifying a high affinity antibody is only part of the challenge.

Successful therapeutics must also possess

Favorable properties such as stability, low aggregation, appropriate viscosity, and manufacturability.

A third thing is data integration.

Many groups are moving beyond using a single data source, and instead combining

sequence structure, biophysical measurements, and experimental data sets to improve prediction accuracy.

А фанали інтер

uh interpretability is becoming increasingly important.

Researchers want models that not only make accurate predictions, but also provide mechanistic insight that help guide experimental design and decision making.

Well, I'd like uh also to see more attention given to uh formulation development.

So predicting how antibodies behave under different formulation conditions, such as different excipients, pH, ionic strength, and concentrations is becoming increasingly important as more therapeutics are designed for high concentration subcutaneous delivery.

Another area is I found there is a uh still relatively limited work on emerging modalities such as bispecific antibodies, antibody drug conjugates, and multi-specific therapeutics.

So these molecules present new challenges that often require models beyond those originally developed for conventional maps.

Uh we think we are entering a very exciting phase.

The first major direction will be multi-model AI models that integrate sequence, structured, experimental, uh biophysical data, molecular simulations, and clinical information into unified predictive frameworks.

These models should provide a much more complete understanding of antibody behavior.

The second trend will be more physics informed AI.

Rather than treating machine learning as a black box, future models will increasingly incorporate

biophysical principles and mechanistic understanding, making predictions both more accurate and more interpretable.

I also expect foundation models specifically trained for antibodies and therapeutic proteins to continue improving, enable more efficient antibody design and optimization.

Showing 41–60 of 67 · page 3 of 4