As peptide therapeutics surge into the mainstream – driven by the success of GLP-1 drugs – analytical scientists are being pushed beyond the comfort zone of small-molecule thinking.
Kelly Broster, Senior Manager of Pharma & Biopharma Market Development and Collaborations at Thermo Fisher Scientific, has spent more than 15 years working at the intersection of mass spectrometry and biopharmaceutical development. With a background spanning cancer research, quantitative proteomics, and biomarker assay development, she now focuses on advancing analytical strategies that can keep pace with increasingly complex modalities.
Here, she explores why peptides demand more than pass/fail metrics – and how labs are adapting their workflows accordingly.
What’s the single biggest analytical misconception you still see teams bring from small-molecule workflows into peptide development?
The biggest misconception is assuming liquid chromatography (LC) purity and intact mass confirmation are sufficient, when peptide therapeutics demand deep structural and microheterogeneity characterization, not just chemical purity.
That assumption comes from small-molecule workflows, where a high reversed-phase high-performance liquid chromatography (RP-HPLC) purity value plus mass confirmation often signals control. But glucagon-like peptide-1 (GLP-1) analogs show why that framework breaks down. Peptides are inherently microheterogeneous. Material can appear “clean” by conventional chromatography while still containing sequence variants, deamidation, oxidation, acylation variants, or other subtle modifications that impact potency, stability, and potentially immunogenicity.
Many of these species are isobaric or closely related, meaning a single purity number cannot fully describe molecular integrity. Peptide programs therefore require orthogonal, structure-informing analytics, particularly high-resolution, accurate-mass mass spectrometry (HRAM MS) coupled with ultra-high-performance liquid chromatography (UHPLC), to resolve, assign, and track variants with confidence.
Ultimately, peptide development isn’t about reporting a purity percentage; it’s about understanding and controlling structural microheterogeneity across the molecule’s lifecycle.
At what point do traditional small-molecule analytical assumptions start to fall short when applied to peptide therapeutics like GLP-1s?
What tends to show its limitations first is confidence in the primary readout, particularly when teams apply small-molecule assumptions to peptides and expect a single chromatogram and a single purity value to describe the molecule adequately.
With peptides, that expectation quickly becomes strained. Multiple structurally related species can co-elute, partially resolve, or sit beneath the main peak, and the primary assay may quantify them without truly identifying them. GLP-1 analogs illustrate this clearly. Beyond classical process impurities, you are managing sequence-level variants, isomerization, epimerization, and side-chain modifications that may be present at low levels yet still influence potency, stability, or immunogenicity.
The second inflection point comes when conventional LC conditions prove insufficient for confident structural assignment. Methods originally optimized for chromatographic appearance are not always compatible with high-confidence mass spectrometry (MS) interrogation. In practice, seemingly tactical decisions, such as selecting formic acid versus difluoroacetic acid to balance separation efficiency with MS sensitivity and mass accuracy, become strategically important much earlier than expected.
Ultimately, the shift occurs when a program must explain what changed and why, not simply whether a batch met a numerical limit. At that stage, workflows built primarily to quantify impurities begin to fall short, because peptide development requires tools capable of resolving and assigning microheterogeneity with structural specificity.
What are the main analytical challenges or bottlenecks for GLP-1s as programs move from early R&D into process development and QC?
As GLP-1 programs transition into process development and quality control (QC), the analytical focus shifts from characterizing a sample deeply once to measuring consistently, trending intelligently, and defending decisions under regulatory scrutiny. One challenge faced by labs is detecting and identifying impurities at very low concentrations in a way that is both structurally confident and operationally scalable. GLP-1s, like many peptides, are prone to a range of impurity types, including deletions, insertions, substitutions, isomerization, and chemical modifications such as oxidation or deamidation. Even small amounts can impact biological activity, narrowing the margin for analytical uncertainty.
A second bottleneck is that the species you most need to understand are not always the easiest to characterize with a single technique. That’s why modern workflows lean heavily on UHPLC coupled to HRAM-MS to combine separation with accurate mass and informative fragmentation, enabling both impurity profiling and structural verification. In practice, this could include deploying complementary MS/MS activation (for example HCD alongside ETD/EThcD) to avoid blind spots for labile modifications and to localize changes with confidence.
Finally, as programs mature, questions around structural heterogeneity extend beyond “what’s in the main peak” to include behaviors such as self-association or aggregation; approaches like size-exclusion chromatography-MS (SEC-MS) native and denaturing conditions can become important to distinguish reversible association from covalent aggregates.
Where do you draw the line between methods that are ideal and methods that are actually deployable?
It is a careful balance of risk and actionability. High-end characterization (multi-activation MS/MS, extensive mapping, deep impurity ID) is essential when you’re:
Establishing the critical quality attributes (CQAs) and understanding degradation/impurity mechanisms
Building a comparability argument
Troubleshooting process excursions where a small shift can have a big biological consequence
Routine QC, however, must be deployable: robust, validated, transferable, and efficient. The goal is not to “dumb it down,” but to translate the depth of characterization into a pragmatic control strategy, typically a focused set of targeted assays that monitor the attributes that truly drive performance.
Where this is heading, and where we are already seeing momentum, is MS moving from being research only into regulated, end-to-end workflows with the right software, compliance tooling, and scalability, so you are not forced into a false choice between ideal data and QC reality.
If you had to name one shift in mindset for peptide analytics, what is it?
The key mindset shift is moving from a purity-centric view to a structure-and-heterogeneity-centric view. For example, with peptides, microheterogeneity is not a rare exception – it is a fundamental characteristic of the product that must be measured, understood, and controlled. Today, GLP-1s have made that very visible because the impurity landscape can include subtle sequence variants and chemical modifications that are easy to overlook if you treat the molecule like a slightly larger small molecule.
Practically, that shift changes how teams design their control strategies. You do not just ask, “Is it pure enough?” you ask, “Do we have orthogonal evidence that confirms identity, assigns impurities, and protects the CQAs that relate to safety and efficacy?” That is exactly where HRAM MS has proved so valuable: it provides molecular level clarity that supports better decisions across the lifecycle, not only at the characterization stage.
What new analytical questions do oral peptides force teams to answer that injectables didn’t?
Unlike injectables, oral peptides force teams to interrogate a much harsher, more variable “delivery environment” within the human body, and to connect product quality to performance through that journey.
For injectable peptides, you largely bypass the gastrointestinal tract, so exposure is typically dominated by distribution and clearance. With oral peptides, the GI environment introduces major, patient-to-patient variability and multiple new failure points (enzymatic degradation, chemical instability, permeability limits, formulation effects), which can make systemic exposure lower, more variable, and harder to interpret analytically.
Analytically, that tends to introduce questions like:
What exactly is the intact fraction after exposure to GI-like conditions, and what are the dominant degradation pathways and fragments?
How do formulation components (protectants, permeation enhancers, coatings) shift the impurity/degradant profile over time and under stress?
How do we measure extremely low levels reliably, when bioavailability is limited and the clinically relevant exposure window can be narrow?
In other words, oral delivery doesn’t just add pharmacokinetic complexity; it adds a new set of product-related species that must be detected, identified, and applied with confidence.
Looking ahead, what do you expect will be the biggest change in peptide analytics?
Over the next 3–5 years, I expect the biggest change will be the continued shift of MS from being perceived as high-end characterization to being integrated more routinely into end-to-end, regulated workflows. The direction of travel is clear: developers want analytical strategies that don’t need to be reinvented at every stage, and the technology ecosystem (instrumentation, informatics, and compliance tooling) is increasingly designed to move from research labs into regulated environments without losing data integrity or scalability.
At the same time, the complexity of peptide modalities and the pace of development are pushing analytics to become more sensitive to low-level species while remaining operationally practical. Capabilities that improve signal-to-noise and help teams confidently surface low-abundance features – paired with software that can translate complex data into consistent reporting – will be central to preventing “measurement lag” as peptide pipelines expand.
Which peptide or modality trend are you most excited about right now – and what analytical capability will matter most to keep that wave from outpacing measurement?
I’m excited by the rapid evolution from “single agonist” peptides into more complex GLP-1-class molecules and candidates, including the diversity we’re already profiling across liraglutide, semaglutide, tirzepatide, and retatrutide.
As complexity increases, the analytical capability that matters most is confident impurity identification and structural verification at low abundance, especially for modifications and variants that are biologically consequential but analytically easy to miss. That means coupling strong separations to HRAM-MS and using the right fragmentation tools so you can localize modifications and distinguish closely related species.
Zooming out: what do GLP-1s reveal about the future of analytical chemistry more broadly?
GLP-1s are a very visible proof point that the future of analytical chemistry is not just about measuring more, but about measuring the right things, earlier, with decision-grade certainty, and then translating those insights into scalable control. GLP-1s highlight several broader truths. Therapeutics are becoming increasingly structurally nuanced, and the analytical burden grows alongside that complexity. Regulatory expectations track scientific reality: when small changes can affect safety or efficacy, methods must be capable of detecting and explaining those changes. Finally, the most impactful analytics are those that can move seamlessly from expert level characterization into repeatable, compliant workflows, supporting development timelines and manufacturing confidence, not as an afterthought, but as foundational infrastructure.
Newsletters
Receive the latest pharmaceutical news, personalities, education, and career development – weekly to your inbox.
