Bo Pang is a Senior Applications Development Manager at Carterra, where he develops and validates high-throughput SPR assays and customer-facing workflows.
CREDIT: CARTERRA
Surface plasmon resonance (SPR) biosensors have become a standard tool in pharmaceutical and biotechnology research, prized for their ability to monitor molecular interactions in real time without the need for labels. Historically, SPR was used primarily for characterizing macromolecular interactions, but its combination of high sensitivity, low sample consumption, and kinetic insight has made it increasingly relevant across drug discovery — from target identification to lead optimization.
Advances in high-throughput SPR are now expanding its role further, enabling researchers to collect kinetic data at a scale previously impractical and to integrate that information into AI-driven discovery workflows. To explore how these advances are reshaping experimental design and early-stage decision-making, DDN spoke with Bo Pang, Senior Applications Development Manager at Carterra. The conversation follows the company’s recent launch of the Carterra Vega, a 48-channel high-throughput SPR system designed to bring label-free binding measurements earlier into the drug discovery workflow.
Why has SPR traditionally been pushed into secondary screening rather than used at the front end of discovery?
Throughput and complexity. Traditional SPR is information-rich but slow due to low throughput and its serial nature. Early screening in drug discovery prioritizes speed and breadth, so SPR is usually considered as a “validation” assay after hits are identified.
What risks are introduced when early screening relies heavily on labeled or proxy assays?
Relying on labeled or proxy assays early in discovery can introduce both measurement and mechanistic bias. The label or reporting system, while necessary to generate a signal, can itself interact with compounds, biologicals, or the target, potentially creating assay artifacts. In addition, because these assays typically provide only a single endpoint readout, they offer limited insight into real-time target engagement and binding kinetics compared with direct, label-free measurements.
Does increasing channel count actually reduce experimental compromise, or just accelerate existing workflows?
Increasing channel count can reduce experimental compromise, not just accelerate existing workflows. With a multichannel many-on-few SPR system, researchers can: (1) include more replicates and controls without difficult trade-offs regarding timelines; (2) run parallel kinetics and test multiple targets or constructs within the same assay; and (3) enable assays for challenging proteins with short half-lives by collecting high-quality data within a tighter time window.
At what point does increased throughput fundamentally change how experiments are designed, rather than just how fast they run?
Higher SPR throughput allows researchers to apply the same rigor to screening that was previously reserved for detailed characterization. For example, multi-concentration, reference-corrected binding experiments can become standard in library screening, generating kinetic parameters (kon and koff) rather than a single endpoint response. While traditional 8-channel platforms screen roughly 2,300 compounds per day, the Carterra Vega platform is the first and only higher-throughput system that can handle entire backbone libraries of more than 20,000 compounds daily, enabling biopharma teams to advance many more projects throughout the year.
Are there specific workflows or biological questions that were previously impractical or too slow to attempt with SPR?
Yes, kinetics-first screening and ranking. Most early assays emphasize equilibrium or endpoint readouts, such as IC50 values, which reflect the concentration required to inhibit a target by 50 percent, or KD, a measure of binding affinity at equilibrium. However, in vivo performance is often influenced by binding kinetics in the human body, which is a complex and dynamic system. SPR is the gold standard for measuring real-time association and dissociation kinetics, but traditional workflows were limited by throughput. Higher-throughput SPR now makes it feasible to collect kinetic parameters earlier and at scale.
What types of binding data are most lacking in current AI-driven discovery efforts?
Binding kinetics data, especially real-time kon and koff. The association rate (kon) describes how quickly a molecule binds to its target, while the dissociation rate (koff) reflects how long that interaction persists once formed. Together, these parameters determine not just whether a molecule binds, but how it engages its target over time.
This distinction matters because in vivo efficacy is frequently influenced more by binding dynamics than by equilibrium affinity alone. Two compounds with similar KD values can behave very differently in a biological system depending on how rapidly they associate with, or dissociate from, their target.
Most AI pipelines are trained on widely available endpoint data like IC50 or KD, but those don’t capture how targets are engaged in the human body over time. To model real in vivo behavior more accurately, AI-driven discovery efforts require large, high-quality, benchmarked kinetic datasets that reflect how interactions evolve over time.
How different are the technical and experimental requirements for screening small molecules versus large molecules on SPR, and where do traditional systems struggle most?
Small- versus large-molecule SPR have different pain points. Small molecules are limited by low signal (due to their low molecular weight), high sensitivity to dimethyl sulfoxide (DMSO) or bulk refractive index (RI) changes, and a higher risk of nonspecific or sticky surface binding. Large molecules produce stronger signals, but often present challenges related to avidity, multivalency, and orientation on the sensor surface. Traditional systems struggle most with the throughput versus data-quality trade-off, because designs that prioritize speed often compromise measurement fidelity.
Looking ahead, how do you see advances in high-throughput, label-free screening changing the way discovery teams make early-stage decisions?
The most important task in early drug discovery is to select and design molecules with favorable properties that translate into later stages, including cellular assays, animal studies, and ultimately clinical trials. High-throughput, label-free screening across multiple targets can generate information-rich, physiologically relevant binding data, enabling data-driven decisions early and de-risking downstream time and resource costs.
