Peptide SAR Analysis: Toward a Toolbox for a Growing Modality

Peptide therapeutics are gaining ground in modern drug discovery, with expanding pipelines across metabolic disease, oncology, cardiovascular disease and beyond. However, the analytical demands they place on peptide medicinal chemistry teams differ from those of small-molecule programs, and no widely adopted tools for peptide Structure-Activity Relationship (SAR) analysis have emerged to address them.

This article examines the current tooling gap, the challenges peptide chemists face as a result, and the improvements that dedicated tools could bring to peptide medicinal chemistry workflows.

A fragmented tooling landscape for peptide SAR

Peptide medicinal chemistry has matured considerably over the past decade; however, the software environment supporting it has not kept pace.

Multiple solutions for peptide drug discovery have emerged in recent years, yet only a few focus on SAR, such as the open-source packages PepSeA by Merck & Co. and PepFuNN by Novo Nordisk. These are valuable contributions that have advanced what is possible. By design, though, they are code-based libraries built for computational scientists. Each addresses part of the workflow rather than spanning data handling, analysis, and reporting within a single environment integrated into a project team's daily tools. Solutions like these can suit specialist groups, but they are harder to adopt broadly across teams. As a result, most of the work remains without dedicated support for peptide SAR analysis.

In their absence, peptide chemists assemble their analyses with the tools at hand: Excel for tracking analogues and assay results, general-purpose chemical drawing software for visualizing structures, and in-house scripts to connect the pieces. Taken together, this setup wasn’t designed for peptide-specific data or for how peptide SAR is analyzed.

The limitations of this setup are felt at every stage of the peptide SAR workflow, from data access and handling to reporting. They shape which analyses peptide teams can carry out, how much confidence they can place in the results, and ultimately how well-informed the design decisions are.

Where the gap shows in peptide SAR workflows

The cost of working without dedicated tools concentrates in three areas of the peptide SAR workflow: handling the data, analyzing SAR, and producing reports.


During our event Discngine Meetup Vol. 5 on “Peptide discovery - Pushing boundaries with innovative science and technology”, our speakers addressed the limitations of applying tools designed for other modalities to peptide medicinal chemistry.

Watch the replay to hear the full discussion.


Handling peptide data

Peptide datasets don’t fit naturally into a spreadsheet, which treats peptides as text rather than chemical entities. For a linear peptide built from natural residues, a one-letter sequence captures most of the molecule. Even then, terminal modifications such as N-acetylation or C-terminal amidation are easily lost.

Most peptide programs, however, go well beyond natural residues, and this is where the one-letter code breaks down. Cyclic backbones, disulfide bridges and staples, side-chain modifications, and custom monomers can’t be expressed in it. Yet these are precisely the modifications peptide programs depend on to improve stability, permeability, target engagement, and all the other properties that make a molecule into a drug.

Illustrative cyclic peptide: Ac-[Cys-Aib-Phe-D-Trp-Lys(octanoyl)-Cys]-NH₂, disulfide-bridged. Termini, the bridge, the lipid and Aib (“X”) are absent from the string.

Adding structure images to the sheet does not solve this. A structure pasted into a cell can be seen, but not searched, compared, or used in calculations. Stereochemistry is another clear example. In one-letter code, D-amino acids are written in lowercase: D-Trp is “w”. If autocorrect or a copy-paste changes it to “W”, the sequence looks the same but describes a different molecule. Non-natural residues are usually entered under custom codes that don’t indicate their configuration, so they may not even be recorded at all.

The sequence itself is just as fragile in a hand-built tracking sheet. A single deletion can require manual renumbering across all dependent columns. Inconsistent monomer codes, extra whitespace, or mistyped modifications pass unnoticed through the dataset and surface only later, once they have already shaped downstream calculations. As analogue series grow, the spreadsheet becomes harder to maintain and harder to trust as a reference for the rest of the workflow.

Analyzing peptide SAR

Beyond data handling, extracting SAR insights from datasets is also challenging.

Single-position scans are a common starting point, but as optimization progresses, most analogues carry modifications at two or more positions, and their effects are often not additive: two changes that each improve potency may cancel out when combined or reinforce each other beyond what either suggests.

Identifying these effects means comparing analogues position by position, which only works if each position refers to the same residue in every analogue. Adding or removing a residue shifts every position after it, and cyclic formats, bridges, and custom monomers make those positions even harder to match. Analogues must therefore be aligned to a reference peptide, and that reference changes over a project: first the native peptide, later a competitor compound or the most recent lead. In a spreadsheet, this alignment is done manually and redone with each design round.

Generating peptide SAR reports

The final challenge is to communicate the SAR in a form that supports project decisions.

The SAR tables and slides prepared for project meetings are often the primary way findings are shared across teams and reviewed by decision-makers.

Producing those outputs is rarely a simple export step. SAR reporting already requires considerable manual effort for small molecules, despite established tools. For peptides, the effort is greater: analogues differ in length and at several positions at once, full structures are too large to read on a slide, and datasets keep growing as new results accumulate. Each new round of results means updating or rebuilding what was shared before.

The work is largely manual: redrawing structures, copying values, and reformatting tables. It is time-consuming and prone to error, and as slides are revised and circulated, team members may end up discussing different versions of the same SAR.


Manual SAR reporting is a well-documented issue in drug discovery, as discussed in our previous article, “The Hidden Costs of Manual SAR Reporting in Drug Discovery.”

The challenge persists for small molecules despite established tools and is greater still for peptides, whose analogues differ in length and at several positions at once.


What a purpose-built approach could change

A purpose-built approach wouldn’t make peptide SAR simpler, but it would keep the chemistry intact, bring the whole series into view, and let more of the available data inform each decision.

For example, Biron and co-workers synthesized the complete N-methylated library of the Veber-Hirschmann peptide cyclo(-PFwKTF-), a potent somatostatin analogue that is not orally available: 30 analogues covering every combination of N-methylation. Only 7 retained affinity comparable to the parent, and analogues carrying the same number of N-methyl groups behaved very differently depending on which positions were modified. Of those 7, only one, methylated at 3 positions, was absorbed after oral dosing, reaching about 10% oral bioavailability in rats.

That is just an example of a library that could be analyzed manually: all 30 analogues share one sequence, so every position matches across the set. Normally, a lead optimization program has no such advantage, since residues are added and removed over successive design rounds and the positions no longer correspond. The comparison has to be built before it can be read, and that’s where today’s tools fall short.

A dedicated environment would change several aspects, including:

  • Analogues would be aligned to a chosen reference (that can change). Once registered as chemical entities, in an open notation such as HELM, they can be aligned to that reference and realigned as it changes, even when they differ in length or modifications.

  • The analysis would rest on the entire series rather than the most recent round. Every compound produced to date, including inactive ones, can then be examined in a single view. Inactive analogues are often left aside, since one that lost activity rarely earns a place in a hand-built table, yet they show which positions don’t tolerate change.

  • Combinations would become directly visible. Reading potency, stability, and permeability against the same aligned positions shows which modifications add up and which don’t, as in the Biron example, but across larger and more structurally diverse series.

The benefit of a dedicated environment is not only in what becomes visible. When preparing an analysis requires less manual work, it becomes practical to explore more questions, compare more hypotheses, and revisit earlier assumptions as new data becomes available.

Conclusion

Peptide therapeutics are shaping the treatment of major disease areas, and those achievements have been reached mainly through the strength of the science and the persistence of the teams involved, rather than through the maturity of the tools supporting them.

Many peptide teams continue to operate with software built for other purposes or other modalities, and much of the cost shows up where it is hardest to recover: in the time spent on preparation rather than interpretation, in the portion of the data that never makes it into the analysis at all, and in design opportunities that go unnoticed because a table makes them hard to see.

What the field needs is an environment designed specifically for this modality: one that recognizes peptide-specific data and supports the way SAR is read across an analogue series, from the first analogue to the most recent. Closing that gap would not only ease the daily work of peptide medicinal chemists but also strengthen the analytical foundation on which peptide drug discovery continues to grow.

What such an environment should look like in practice is a question the field is currently addressing.

Summary

Peptide medicinal chemistry lacks dedicated tools for peptide SAR analysis, impacting data handling, analysis, and reporting. This gap limits efficiency and confidence in drug discovery workflows.

  • Peptide teams rely on spreadsheets and general-purpose SAR software, where non-natural amino acids, chemical modifications, terminal groups, bridges, stereochemistry, and more are easily lost, and analogues are hard to align and compare as series diversify.

  • A purpose-built solution would keep the chemistry intact, bring the whole series into view, including inactive analogues, and show how modifications combine across positions.

  • Closing this tooling gap would save time spent on data preparation and strengthen the foundation for advancing peptide therapeutics.

 
Next
Next

Ideation Analytics 3.0: Enterprise-level collaboration and peptide SAR capabilities