Welcome back to Research Weekly.
This week, the FDA peptide story moves closer to the committee room. The agency has updated the schedule for next week’s Pharmacy Compounding Advisory Committee meeting, the public-comment deadline is approaching, and the final meeting materials are beginning to take shape.
We’re also looking at a separate research trend that could have a much longer-term impact: the growing use of artificial intelligence and automated screening to design more complex cyclic peptides.
As always, the goal here is simple: useful research context, clear takeaways, and less noise.
FDA Peptide Meeting is Now One Week Away
The FDA’s Pharmacy Compounding Advisory Committee is scheduled to meet on July 23 and 24 to consider seven peptide groups and their free-base and acetate forms for possible inclusion on the 503A Bulks List.
The substances scheduled for discussion are:
July 23
BPC-157
KPV
TB-500
MOTS-C
July 24
Emideltide, also known as DSIP
Semax
Epitalon
FDA staff has proposed that none of the reviewed free-base or acetate forms be added to the 503A Bulks List. The committee will review the agency’s analysis, hear presentations, consider public input, and vote on recommendations for each substance. Those recommendations are advisory and will not represent the FDA’s final determination.
What Changed This Week?
On July 14, the FDA updated the meeting schedule.
The July 23 session is now expected to continue until 6:20 p.m. Eastern Time, rather than ending at 4:30 p.m. The July 24 session is now scheduled to end at 4:15 p.m., rather than 3:50 p.m.
The agency also changed the scheduled times for oral presentations from members of the public. All other meeting information remains unchanged.
Public Comments Close July 22
The public-comment docket remains open until 11:59 p.m. Eastern Time on July 22.
Comments submitted by July 9 are expected to be provided directly to the committee. Comments received after July 9 but before the final deadline may still be considered by the FDA as part of its broader review.
What We’re Watching Next
The biggest questions next week will be:
Whether the committee agrees with FDA staff on every substance
Whether members distinguish between free-base and acetate forms
How the committee evaluates limited human evidence
How much weight is placed on preclinical studies
Discussion surrounding identity, purity, stability, and manufacturing controls
Whether public presentations introduce meaningful new evidence
How the committee handles compounds whose original nominations were withdrawn.
The FDA briefing documents repeatedly emphasize that a commonly used peptide name does not always identify one clearly characterized chemical substance.
That means questions involving exact chemical form, reference standards, analytical methods, stability, and batch documentation may be just as important as the biological research being discussed.
We will publish a clear summary after the meeting, including the committee’s votes and the major arguments raised during the discussion.
AI is Moving Deeper into Cyclic-Peptide Design
Most familiar peptide sequences are represented as linear chains of amino acids.
Cyclic peptides are different. Their structures are chemically linked into rings, which can restrict how the molecule moves and folds. That additional structural control can improve properties such as target binding and resistance to degradation, depending on the peptide and linkage involved.
The challenge is that researchers must determine more than the amino-acid sequence. They must also decide:
Where the peptide should be cyclized
Which residues should be linked
What type of chemical linkage should be used
Whether the resulting structure maintains its target interaction
How the cyclization affects stability, solubility, and other properties
A recent preprint called APCyc describes an artificial-intelligence framework designed specifically for this problem.
Rather than treating cyclic peptides like ordinary linear sequences, the system represents cyclization sites and linkage types directly. It then attempts to generate target-specific cyclic-peptide candidates while optimizing multiple physicochemical properties at the same time.
The authors report that the model learned target-dependent cyclization preferences and could guide peptide generation toward selected property goals. The research is currently a preprint and has not yet completed peer review, so its findings should be viewed as preliminary.
Automated Screening is Advancing Too
Computational generation is only one side of the trend.
Researchers also continue to improve laboratory screening systems capable of searching extremely large libraries of nonstandard cyclic peptides.
A recent study in ACS Chemical Biology used the RaPID screening platform to identify a de novo macrocyclic peptide that inhibited the related targets ROR1 and ROR2. The work illustrates how large-library screening can uncover cyclic structures that may be difficult to design through conventional sequence selection alone.
Why it Matters
Peptide research is increasingly becoming structure-aware, rather than focusing only on the written amino-acid sequence.
Researchers are paying closer attention to:
Three-dimensional conformation
Cyclization and linkage chemistry
Chemical modifications
Target-specific binding
Stability and degradation
Membrane permeability
Reproducibility and characterization
AI and automated screening may help researchers explore a much larger design space, but they do not remove the need for synthesis, analytical confirmation, laboratory validation, and careful documentation.
In many ways, this connects directly to the questions raised in the FDA briefing documents: the name of a peptide is only the beginning. Its exact structure, chemical form, manufacturing process, and analytical identity all matter.
0 comments