Thirty steps made it. One decides the purity.

This is the written version of Cheap Peptides Aren’t Cheap For The Reason You Think — watch it instead if you’d rather.

A peptide is not poured, cast or grown. It is built, one amino acid at a time, on a bead. You anchor the first amino acid to a solid support, couple the next one on, wash the bead clean, uncap it so the next coupling can happen, and repeat until the chain is the length you wanted. Then you cut the finished chain off the bead. That is solid-phase peptide synthesis, and I think of it the way I would think of any process train with dozens of sequential unit operations: every single coupling is a separate, independent chance for that one step to fail.

That framing matters, because most people hear a purity number and picture a defect rate — a percentage of bad units on an otherwise fine line. That is not what is happening. The chain is assembled step by step, and the steps do not average out. They compound.

Yields multiply, they do not add

Semaglutide is thirty-one amino acids long, which means thirty separate coupling steps. Run every one of those thirty steps at a genuinely excellent ninety-nine per cent yield — better than most real-world chemistry manages — and you do not finish at ninety-nine per cent. You finish at about seventy-four. Multiply 0.99 by itself thirty times and that is the number you get. Slip to a still-respectable ninety-eight per cent per step and you are down to about fifty-four per cent. Nothing has gone wrong at any individual step in either case. That is simply what excellent looks like, done thirty times over, because yields on a sequential process compound rather than add.

The other quarter is not dirt

So what is sitting in that missing twenty-six per cent, or worse? Not contamination in the way people usually picture it. It is near-misses: chains where one amino acid never coupled, so the finished molecule is a residue short. Chains that stopped growing early. Chains where a protecting group never fully came off.

• Deletion — a residue that never coupled.
• Truncation — a chain that stopped early.
• Incomplete deprotection — a protecting group left in place.

None of that is a rounding error. A peer-reviewed review of industrial peptide purification opens from a real manufacturing batch of calcitonin sitting at forty-six per cent crude purity. That is what genuinely comes off the machine before anyone has separated anything. And the same review makes the awkward point plainly: something missing one amino acid out of thirty-one is chemically almost identical to the molecule you wanted. Near-identical molecules move through a purification column at nearly the same speed, which means they arrive at the collector at nearly the same moment as the thing you were trying to isolate.

The dial: purity against yield

This is the hinge of the whole story. You push the crude mixture down a chromatography column and different molecules travel through it at different speeds. Your target comes out somewhere in the middle. The near-misses come out either side of it, overlapping at the edges, like shoulders either side of a central peak. Somebody then has to decide, physically, when to start collecting and when to stop.

Narrow that collection window and you get purity, because you are throwing away the overlapping shoulders along with everything in them. Widen it and you keep far more of the batch, but the near-misses ride in with it. The review states this as a general limitation of the method rather than a one-off: in batch chromatography it is practically impossible to reach both a high purity and a high yield at the same time on a single column. It is a dial, not a switch, and somebody has to choose where to sit on it.

The same review gives a concrete case. Purifying icatibant past ninety-nine per cent purity on a single batch column came out at a recovery of about twelve per cent. Make a kilogram and you keep roughly a hundred and twenty grams; the rest is the price of that purity number. Put that beside another figure from the same source — downstream purification runs somewhere between fifty and seventy per cent of the total cost of making the peptide in the first place — and the decision about where to set that collection window becomes, in cost terms, the single most expensive decision in the building.

Which is why the economics point one way. Widen the window a little and yield multiplies while the near-misses ride along, and cost per gram falls. I want to be precise about what that is and is not. The yield-versus-purity trade-off is a documented, peer-reviewed limitation of the method itself. Anybody actually stating that they widened a cut to save money is not something I have a source for, so I am not writing it as though I did. It is an incentive built into the economics, not an accusation. And for what it is worth, I take no money from anybody who sells peptides, so I have no reason to shade this either way.

A registered plant is not an approved one

There is a floor under all of this, and it is paperwork. An FDA warning letter, reference 711330, issued on 11 September 2025, set out findings against an ingredient manufacturer that was registered with the FDA. Registration is a filed form, not an approval, and that distinction is worth keeping exact rather than letting it blur into “FDA-approved”, which is not what it means.

The letter found the firm had carried out no process validation for the ingredient — no proof the process would reliably give you the same thing twice. It also found a failure to run an identity test on incoming raw materials, meaning nobody confirmed the amino acids arriving at the gate were the amino acids they were meant to be. The firm was required to commit to always conducting at least one specific identity test for every incoming component lot. What struck me most is how the agency got there: from a records request. It did not need to walk onto the site to find any of it.

Why the leftover matters

One more thing, because it is mostly the right molecule sounds reassuring and should not, on its own. A study published in Frontiers in Immunology in December 2025, with authors from a vaccine-design group alongside co-authors from the FDA’s Center for Drug Evaluation and Research, tested peptide-related impurities found in generic teriparatide products three separate, orthogonal ways — and identified several of them as more immunogenic than the drug itself. The near-miss sitting in the shoulders of that chromatogram is not inert filler. It can be the part a person’s immune system actually notices.

The bottom line

A purity number is not a fixed property of a molecule, the way a melting point is. It is a decision somebody made about where to cut a collection window, on a process where the arithmetic already compounds against you thirty times before purification even starts.

In batch chromatography, it is practically impossible to reach both a high purity and a high yield at the same time.

When people ask me why cheap peptides are cheap, they usually expect a story about criminals. Most of the time it is a story about a collection window, and about how much of the total cost of the thing sits in the decision to narrow or widen it.

Had a certificate sent to you? I read them — peptidecorner.net/coa-read

If you want the free rundown on what a certificate can and cannot tell you on its own, it is one of the three free guides on the site. And if the question you actually have is which of four systems your own vial came out of before any of this arithmetic applied, that is a separate piece. This one is about what happens inside whichever machine you land on, and why the last step decides the number on the certificate.

Sources

1. De Luca et al., Molecules, 2021 — the review the middle of this video rests on. In batch chromatography it is practically impossible to reach both high purity and high yield at the same time. Also the source of the calcitonin crude at 46% purity, the point that impurities differing by a single residue co-elute, the estimate that roughly 50 to 70% of manufacturing cost is downstream processing, and the icatibant case at over 99% purity for around 12% recovery on a single batch column — https://doi.org/10.3390/molecules26154688
2. Mattei et al., Frontiers in Immunology, 8 December 2025 — EpiVax authors with FDA CDER co-authors. Peptide-related impurities identified in generic teriparatide products were tested three separate ways, and the orthogonal approaches identified multiple impurities as more immunogenic than the drug itself — https://doi.org/10.3389/fimmu.2025.1730346
3. FDA warning letter, reference 711330, issued 11 September 2025 to an ingredient manufacturer registered with the FDA — no process validation for the ingredient, and a failure to run an identity test on incoming raw materials, with the firm required to commit to always conduct at least one specific identity test for each incoming component lot. The agency established all of it from a records request rather than an inspection. Search the reference number on the FDA warning-letter index — https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/compliance-actions-and-activities/warning-letters

Educational and research purposes only — not medical advice. Peptide Corner does not recommend any vendor, source, or dose. Keep safe, keep skeptical.