Mass spectrometry peptide identity: why retention time alone is not enough

What an LC-MS identity test measures

An LC-MS identity test answers one question: does the main peptide in the vial have the expected molecular mass? Liquid chromatography first separates the sample. The liquid then enters the mass spectrometer, where electrospray turns each peptide into ions that carry several extra protons. The instrument measures their mass-to-charge ratio (m/z).

A large peptide shows up as a series of charge states. Semaglutide has the formula C187H291N45O59 and a calculated monoisotopic mass of 4111.1154 Da. With three extra protons it appears near m/z 1371.38, and with four near m/z 1028.79. Software combines the charge states into one intact mass, a step called deconvolution.

The result is compared with the calculated value as an error in parts per million: ppm = (observed − calculated) ÷ calculated × 1,000,000. A high-resolution instrument measures within a few ppm. In a technical report from HALO, a high-resolution system measured semaglutide at 4111.1027 Da against 4111.1154 Da calculated, an error of 3.09 ppm.

Why retention time alone is not enough

Many labs confirm identity by retention time: the sample peak leaves the column at the same time as a reference standard. ICH Q6A says that identification by a single chromatographic retention time “is not regarded as being specific”. It accepts combinations such as HPLC with a diode array spectrum, or HPLC with mass spectrometry. The EMA synthetic peptide guideline, in force since 1 June 2026, recommends at least two orthogonal identity methods.

The reason is simple. Retention time depends on the column, the gradient and the temperature, and related peptides can leave the column close together. On one lab’s published chromatograms we reviewed, two different peptides eluted 0.16 minutes apart on the same method. ICH Q2(R2) asks that an identity method be challenged with closely related compounds to show it gives no false positive.

Retention time can still catch a swap when the wrong peptide elutes far from the expected one. It cannot prove identity when the two elute close together. A mass reading can, because related peptides almost always differ in mass. Tirzepatide and retatrutide, for example, differ by 82 Da. Our guide to blind GLP-1 identification shows how this works for that family.

How peptide testing labs confirm identity today

The table lists the identity evidence each lab publishes on its own site or on the reports we read in September 2026. It describes the method, not the quality of any single result.

Lab	Identity evidence as published
Ethos Analytics	HPLC retention-time concordance (USP <621>) against a certified reference standard
ILS Laboratories	HPLC retention-time matching against a reference standard; LC-MS identity is a custom add-on
PharmLabs	HPLC retention time against a reference standard; no mass spectrometry listed for peptides
TrustPointe Analytics	Retention-time ratio, UV spectral match and a co-elution control on each result
Vanguard Laboratory	Retention time against a reference standard (HPLC-UV) on the reports we read
Accumark Labs	HPLC identity; LC-MS identity listed as in development when we checked in September 2026
Testides	Retention time and UV match unless mass spectrometry is ordered
Krause Analytical	Full LC-MS spectrum on each peptide report; no expected m/z printed
Freedom Diagnostics	LC-MS identity; reports show a mass spectrum image without an expected mass or error
MZ Biolabs	Expected and measured accurate mass, on QTOF instruments with a resolution of about 20,000
SideChain Analytics	LC-MS/MS on a triple-quadrupole instrument, with precursor and fragment transitions
Ozcanium Analytics	LC-MS/MS: monitored and detected m/z for several charge states

Across the market the picture is mixed. In our review of public COAs from 25 labs and one rating site, 12 showed a mass spectrum, and only 3 printed expected and observed mass side by side with an error value.

What a good identity result shows on a COA

A mass result you can check prints:

• the formula used and its source, so the expected mass can be recomputed;
• the expected m/z for each charge state and the observed m/z next to it;
• the error in ppm for each ion, and the intact mass after deconvolution;
• at least two charge states, and the spectrum itself;
• the retention window where the spectrum was taken;
• wording such as “consistent with”, with the identity method named.

You can check the arithmetic yourself. Expected m/z = (monoisotopic mass + z × 1.00728) ÷ z, where z is the charge. Our review found reports with charge states labelled wrongly and one that printed a measured mass identical to the calculated mass, 0.0 ppm. A real measurement almost never lands exactly on the calculated value.

A good report also lists related species by mass. Oxidation adds about 16 Da and deamidation about 1 Da. The HALO report found a +12 Da change at the N-terminal histidine in a compounded semaglutide sample, at a level above that of the reference product. An identity test that prints these shifts tells you more than a pass mark.

What LC-MS cannot tell you

Mass spectrometry proves the mass, not everything about the molecule:

• Same-mass forms. D-amino-acid isomers and isoaspartate forms have the same accurate mass as the parent peptide. A 2022 study of semaglutide’s D-amino-acid isomers needed separation by chromatography to see them. Leucine and isoleucine also share a mass.
• How pure the peptide is. Different molecules ionise with different efficiency, so peak heights in a mass spectrum are not percentages. Purity needs HPLC. See our guide on HPLC peptide purity.
• How much is in the vial. Content is a separate test against a reference standard.
• Names that cover several molecules. “TB-500” and “CJC-1295” each describe more than one molecule. The report should print the exact sequence or formula it tested for.

LC-MS/MS and sequence confirmation

LC-MS/MS adds a fragmentation step. The instrument selects a peptide ion, breaks it, and measures the pieces. Backbone fragments, known as b and y ions, read out stretches of the sequence. A report can then give sequence coverage as a percentage. The EMA guideline lists LC-MS/MS of the intact molecule, peptide mapping, NMR and Edman degradation as ways to confirm a sequence.

A triple-quadrupole instrument often runs in multiple reaction monitoring (MRM). It watches for a chosen precursor ion and one or two chosen fragments. That is a sensitive check for a target molecule, but it only sees the transitions it was set to watch. It shows that the expected molecule is present, and little about what else is there.

How HK Analytical reports identity

Our identity test compares the measured mass of the main peptide with its expected mass. The report prints expected and observed mass with the error in ppm, and the mass spectrum behind it, as our example report shows. For GLP-1 family peptides, blind GLP-1 identification checks the vial against a list of candidates without being told the label. We do not offer sequence confirmation by LC-MS/MS. You can combine identity with purity, content and other tests per sample through a quote.