Most quality discussions about research peptides start at the vial and work backward: what does the HPLC trace show, what mass did the spectrometer see, how much of the powder is actually peptide. Those are the right questions. But they all sit downstream of one fact that almost never appears on a certificate of analysis — how the molecule was made in the first place.
Manufacturing route is not a trivia detail. It determines the menu of impurities a compound can plausibly contain. A chemically synthesized peptide can carry deletion sequences and epimers. It cannot carry host cell proteins, because there was never a host. A recombinantly expressed protein can carry residual host DNA and endotoxin. It cannot carry a piperidine adduct from an Fmoc deprotection step that never happened.
Knowing the route tells you which tests are load-bearing and which are decoration. Here is the taxonomy.
Two constraints decide which route is even possible
Manufacturers don't pick a production route by preference. Two hard constraints usually make the decision for them.
Length. Solid-phase peptide synthesis (SPPS) builds a chain one residue at a time on a resin bead. Every coupling step is very efficient but not perfect, and the small failures compound. In practice, SPPS is the workhorse for peptides up to roughly 50 residues, is workable with effort out to around 70, and beyond about 80 residues recombinant expression is generally the more viable manufacturing path. Chemical ligation strategies — native chemical ligation, developed in the 1990s, joins deprotected fragments via an N-terminal cysteine and a C-terminal thioester — extend the chemical route further, but they add steps and cost.
Modification the ribosome can't do. A ribosome installs L-amino acids from the standard genetic code. If a molecule contains a D-amino acid, a non-proteinogenic residue, or a synthetic side chain, a standard expression host simply cannot build it. Conversely, if a molecule needs complex glycosylation to be the real thing, chemistry alone can't build it either — and neither can E. coli, which doesn't glycosylate.
Those two constraints sort essentially the entire catalog.
Route one: solid-phase chemical synthesis
This is the route for most of what the research market calls a "peptide." Short chains — Ipamorelin at five residues, GHK-Cu at three, BPC-157 at fifteen — sit comfortably inside the SPPS window.
The process is a repeating cycle: deprotect the growing chain's N-terminus, couple the next protected amino acid, wash, repeat. At the end the peptide is cleaved from the resin — in Fmoc chemistry, with a strong acid cocktail plus scavengers — then purified by preparative reversed-phase HPLC and lyophilized. Each step has a characteristic failure mode, and each failure mode is a named impurity class:
- Deletion sequences — a coupling step didn't go to completion, so a fraction of chains are missing a residue. These are close in mass to the target and can co-elute.
- Truncated sequences — chains that stopped growing entirely.
- Insertion sequences — an extra residue added.
- Aspartimide formation — base-catalyzed cyclization at aspartate, notoriously bad at Asp-Gly motifs where glycine offers the least steric hindrance. A 2025 ChemBioChem review catalogs the protecting-group strategies used to suppress it. The consequence matters: aspartimide opens to give α- and β-peptides and racemized product.
- Epimerization / racemization — an L-residue flips to D during coupling.
- Adducts — piperidine, scavenger, or protecting-group residues carried through cleavage.
- Disulfide scrambling — in cysteine-containing peptides, the wrong pairing.
What catches them: RP-HPLC for the separable ones, LC-MS for mass-shifted ones. But note the gap. An epimer has exactly the same mass as the correct peptide. Mass spectrometry is structurally blind to chirality — catching it requires chiral analysis or amino acid analysis after hydrolysis, which is not a standard research-grade COA line item.
Route two: recombinant expression
When a molecule is too long to synthesize, it gets made by a living cell. A gene is inserted into a host — E. coli, yeast, or a mammalian cell line — the host transcribes and translates it, and the product is harvested and purified from the cellular soup.
The catalog's clearest example is HGH 191AA. Human growth hormone is 191 amino acids. That is far past the practical SPPS ceiling, so recombinant expression is the only realistic route — and the compound's own name is a manufacturing artifact. First-generation recombinant hGH made in E. coli retained the initiator methionine, giving a 192-residue methionyl-hGH (somatrem, marketed as Protropin, approved 1985). Later processes yielded the authentic 191-residue sequence. The "191AA" designation is literally a statement about which generation of manufacturing produced the molecule.
Worth correcting a piece of market lore here. The 192-residue form was more immunogenic in clinical use, but a Genentech account of the Protropin development program (Jones, Developments in Biologicals 2002;109:107–118) attributes that immunogenicity not to the N-terminal methionine and not to E. coli protein impurities, but probably to small amounts of growth hormone carrying subtle structural alterations whose nature was never characterized — with manufacturing improvements reducing the response and antibody-associated growth attenuation reported at under 0.1% of patients. "191AA versus 192AA" is a real identity distinction, and a good illustration that route-linked defects can be structural rather than sequence-level. It is not the safety cliff it's sometimes sold as.
The recombinant impurity menu is completely different from the chemical one:
- Host cell proteins (HCPs) — the host's own proteins that survived purification.
- Residual host DNA.
- Endotoxin — if the host is Gram-negative, the host is the endotoxin source, not an incidental contaminant. We covered this distinction in endotoxin testing.
- Misfolding and incorrect disulfide pairing — the sequence can be right while the three-dimensional structure is wrong.
- Glycan variability, for glycosylated products.
What catches them: HCP and host-DNA assays, LAL or recombinant-factor endotoxin testing, peptide mapping to confirm disulfide connectivity, and higher-order structure methods. An HPLC purity percentage says almost nothing about folding.
Route three: extraction from a biological source
The third route is the oldest and the one most people forget exists. Some compounds are still purified out of a natural biological starting material.
The gonadotropins are the standing example. HMG — human menopausal gonadotropin, menotropin — is by definition extracted from the urine of postmenopausal women. HCG is classically purified from the urine of pregnant women, though a recombinant version (choriogonadotropin alfa) also exists; the name alone doesn't tell you which lineage a given preparation came from. Both are heterodimeric glycoproteins, and the glycosylation isn't decoration — sialylation influences circulating half-life and measured biopotency. That's why neither SPPS nor plain bacterial expression is an option.
This route has its own impurity signature: co-purified human proteins. Compositional analysis of a urinary menotropin preparation found gonadotropin content well under total protein, and identified major co-extracted species including leukocyte elastase inhibitor, protein C inhibitor, and zinc-α2-glycoprotein.
And here the analytical paradigm changes entirely. Because the active species is a heterogeneous mixture of glycoforms rather than a single defined molecule, potency isn't reported as a purity percentage at all — it's reported in International Units, defined by bioassay against an international reference standard. The classic FSH method is the Steelman–Pohley rat ovarian-weight augmentation assay (1953), and work such as a 2023 IJMS paper has been pushing toward validated in vitro replacements. By contrast, recombinant FSH can be quantified by protein mass, because it's a defined molecule.
That is the deepest lesson in this whole taxonomy: for an extracted glycoprotein, "98% pure" is not a meaningful claim, and its absence is not a red flag. You are looking for a potency assignment and a reference standard, not a chromatogram percentage. Comparing an HPLC number on a synthetic tripeptide to an IU designation on a urinary gonadotropin is comparing two different sciences.
Structural tells: reading the route off the molecule
You usually can't know your vendor's process. You can often infer the constraint.
A D-amino acid forces a chemical route. SS-31 contains D-arginine and 2',6'-dimethyltyrosine. Adipotide carries a D(KLAKLAK)₂ payload. Dermorphin has D-alanine at position two. None of these can come off a standard ribosome as-is — chemistry is doing the work, so the chemical-route impurity menu applies, chirality blindness included.
Non-proteinogenic residues and synthetic side chains do too. The lipidated incretins carry Aib substitutions and fatty-diacid side chains — neither is something a standard expression host installs, so at minimum a chemical step is involved. Whether a given manufacturer uses full chemical synthesis or a hybrid process is proprietary and not inferable from the vial.
Coordination complexes add an axis. GHK-Cu is a tripeptide plus a copper(II) ion. Peptide purity and copper content are separate measurements; a clean peptide with wrong copper stoichiometry is not the right material.
And some catalog items aren't peptides at all. NAD+, AICAR, and 5-Amino-1MQ are small molecules with small-molecule analytics — a point we made in the metabolic class primer.
What this changes about reading a COA
It changes what counts as a complete document, and completeness is route-dependent.
For a short synthetic peptide, the useful panel is RP-HPLC purity, mass-spec identity, net peptide content, counterion and water content, and ideally a note on chiral purity. Host cell protein testing would be meaningless.
For a recombinant protein, HPLC purity is necessary but nowhere near sufficient. Ask about endotoxin, host cell protein, and folding confirmation.
For an extracted glycoprotein, ask about potency assignment and reference standard, not a purity percentage.
The general rule: a test only has value if it can detect a failure mode that route can actually produce. A COA that reports the same three lines for every product on the shelf isn't characterizing those products — it's using one template. Our Quality & Testing page covers our documentation standards, and the compound library lists sequences and structures for the compounds referenced here.
FAQ
Does a higher HPLC purity number mean a better manufacturing route? No. The number is route-relative. A 99% figure on a synthetic pentapeptide and a 95% figure on a 40-residue peptide represent very different synthetic difficulty, and a purity percentage on a heterogeneous extracted glycoprotein may not be an appropriate metric at all.
Can mass spectrometry tell me how a peptide was made? Only indirectly. It confirms mass, and mass mismatches can point to specific route-linked defects — a +2 Da shift suggesting a reduced disulfide, an adduct mass suggesting incomplete cleavage. But it cannot distinguish D from L residues, which are isobaric, so it cannot detect epimerization.
Why don't research vendors state their manufacturing route? Process detail is generally treated as proprietary, and many vendors are resellers who don't have it. That is precisely why the analytical panel matters: the route is usually invisible to the buyer, so the COA is the only window onto it.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.