Why Two Labs Report Different Purity for the Same Vial

Two laboratories can test the same vial and report different purity, and both can be correct. Chromatographic purity is not a fixed property of the molecule. It is the output of a method applied to a sample, and it changes when the method changes. The number answers the question: of the material this method could separate and this detector could see, what share is the main compound.

That sentence carries most of the explanation. What follows takes the method apart, one variable at a time, and then separates the differences that are expected from the ones that are not.

Purity is a property of the method as much as the molecule

Peptide purity is usually measured by reversed-phase high-performance liquid chromatography with ultraviolet detection. The sample is injected onto a column, a gradient of increasing organic solvent moves the components through it, and a detector records what leaves the column over time. The result is a chromatogram. Purity is reported as area percent: the area of the main peak divided by the total area of all peaks.

Every step of that chain contains choices. Each choice affects which impurities are separated from the main peak, which are seen at all, and how their areas are drawn. Our note on peak area and purity covers the calculation itself. This article covers why the inputs to it differ between laboratories.

A purity figure with no method beside it is incomplete. Two figures with different methods behind them are not measurements of the same thing, even when the vial is the same.

The column: chemistry, pore size and particle size

The column is where separation happens, and its packing determines how a peptide interacts with it.

Bonded phase. The most common reversed-phase packings are C18, C8 and C4. A C18 phase is more hydrophobic and generally retains peptides more strongly. A C8 or C4 phase retains them less. Different phases can change how closely two species elute, and sometimes the order in which they elute.

Pore size. Larger molecules need pores wide enough to enter. Analytical methods for peptides commonly use wide-pore packings for larger peptides and narrower pores for smaller ones, and the choice is made to suit the size of the molecule. A pore size that suits one peptide may serve another less well.

Particle size and column length. Smaller particles and longer columns generally give sharper peaks and better separation. A closely eluting impurity that is a shoulder on one column can be a distinct peak on another. Under the first column it is folded into the main peak. Under the second it is counted separately.

The consequence is direct. A method that fails to separate an impurity from the main peak reports a higher purity, not because the impurity is absent, but because it is hidden inside the main peak.

Mobile phase and gradient

The solvents and the way they change over the run also move the result.

Ion-pairing agent. Trifluoroacetic acid is widely used in peptide methods. It controls pH, pairs with positively charged groups on the peptide to increase retention, and suppresses unwanted interactions with the silica surface. A method using formic acid in place of trifluoroacetic acid will behave differently, with different peak shapes and sometimes different spacing between species. Formic acid is often chosen when the same method must be compatible with mass spectrometry, because trifluoroacetic acid can suppress ionisation.

Gradient slope. A shallow gradient spreads peaks apart. A steep gradient compresses them. A steep gradient finishes sooner and can bring a closely related impurity together with the main peak. A shallow gradient takes longer and can separate it. Slope mainly alters the distance between peaks, so a change in slope can move an impurity from hidden to visible without changing the sample at all.

Flow rate and column temperature. Both change retention and peak width. Temperature in particular can alter selectivity for peptides, and a method run at a different temperature may not reproduce another laboratory's spacing exactly.

Injection load. Injecting too much sample broadens the main peak and can swallow nearby impurities. Injecting too little can leave small peaks below the level the detector can distinguish from noise. Both change the area percent.

For a closer look at how these variables set a peak's position, see retention time explained.

Detection wavelength

Area percent is built from detector response, and the detector does not respond equally to everything.

The peptide bond absorbs strongly in the low ultraviolet, which is why peptide methods commonly detect near 214 nanometres. At that wavelength nearly every peptide-related species gives a signal, including deletion sequences and truncations, because they all contain peptide bonds.

A wavelength near 280 nanometres behaves differently. It responds mainly to aromatic residues, tryptophan and tyrosine in particular. An impurity lacking those residues may produce little or no peak at 280 nanometres, so the main peak looks purer than it is. Two laboratories reading the same vial at 214 and 280 nanometres can therefore report different purity from the same physical injection.

There is a related limit. Area percent assumes that each species contributes area in proportion to its amount. That holds only if every species responds to the detector in the same way. In practice response differs between species, so area percent is an approximation of composition, not a mass fraction. Anything that does not absorb at the chosen wavelength is not counted at all. The page on how peptide purity and identity are measured sets out what that excludes.

Integration: where the baseline goes

After the instrument records the trace, someone or something has to decide where each peak starts and ends. This step is less visible than the column or the gradient, and it moves the number just as much.

Baseline placement. The area of a peak is the area between its trace and a baseline. Draw the baseline a little higher and the peak is smaller. Draw it lower and the peak is larger. In a gradient run the baseline often drifts, because the mobile phase absorbs differently as its composition changes. How the baseline is drawn through that drift affects small peaks far more than large ones.

Drop and valley methods. When two peaks are not fully separated, the software has to divide the shared area. A perpendicular drop from the valley between them to the baseline is one approach. Other approaches follow the valley more closely or treat the small peak as sitting on the tail of the large one. Integration of partly resolved peaks is a known source of error, and the direction and size of the error depend on the relative size, order and width of the two peaks. The same pair of peaks can give a different split of area under a different rule.

Automatic and manual integration. Automatic integration applies fixed parameters, such as a minimum peak height and a noise threshold. Manual integration lets an analyst adjust the result. Both are legitimate. A laboratory should document which was used and why, because manual changes are a place where a result can move without the sample changing.

Reporting threshold. Many methods ignore peaks below a stated threshold. A laboratory that excludes small peaks reports a higher purity than one that includes them, with identical raw data.

The sample, the preparation and the reference standard

Not every difference comes from the instrument. Some comes from what was injected.

Sample preparation. Solvent, concentration and time between dissolving and injecting can all change the sample. Peptides containing methionine, cysteine or tryptophan can oxidise in solution, and others can deamidate. A sample left standing before injection can show more impurity than the same sample run immediately.

The material itself. Two laboratories may not test the same material even when they believe they do. A vial opened, resealed, shipped or stored differently can have changed. Vials from the same batch can also differ, which is covered in single vial versus batch testing.

The reference standard. Identity is usually confirmed by comparing retention behaviour to a reference standard. The quality and characterisation of that standard sets how much weight the comparison carries, as described in reference standards in peptide testing. Two laboratories using different standards, or none, are anchoring their results to different things.

Which differences are normal and which are not

A difference between two results is not itself a finding. What matters is whether it has an explanation.

Differences that are expected:

• Both results are high and close, and the methods differ in column, gradient or wavelength.
• One method resolves a shoulder that the other folds into the main peak.
• The two laboratories apply different reporting thresholds for small peaks.
• The samples were prepared, stored or tested at different times.
• Detection was at different wavelengths.

Differences that warrant a closer look:

• A large gap between the two figures with no difference in method that could account for it.
• A chromatogram that does not match the stated figure, for example a visible impurity peak that the reported purity does not reflect.
• A result with no method, wavelength or integration approach stated.
• Integration that was changed manually with no recorded reason.
• A purity figure reported with no chromatogram to inspect.
• A figure that cannot be confirmed at the issuing laboratory.

The difference between the two lists is traceability. In the first, the cause can be found by comparing documents. In the second, it cannot.

How to compare two results properly

A comparison is only meaningful when the same questions are put to both.

• What was the method? Column, mobile phase, gradient and detection wavelength.
• What was injected? The same batch, the same vial, and how it was prepared.
• How were peaks integrated? Automatic or manual, and the reporting threshold.
• What is the chromatogram? The trace, not only the figure.
• What was the identity evidence? Mass confirmation, and the reference used. See HPLC versus mass spectrometry for what each can and cannot show.

Without those, two numbers can only be compared as numbers, which is the comparison that misleads.

What a purity figure does not settle

Even when two laboratories agree, the figure has the same limits as any other. It describes the composition of the sample submitted, as seen by one method. It is not net peptide content, which is a separate measurement of how much of the powder by weight is peptide. It is not a statement about sterility, endotoxin or safety. Purity is composition, not safety. And a result describes the material that was tested, not every vial from the same source.

Novagen is a third-party analytical laboratory. It reports what was measured in the sample submitted, and it states the method beside the result. Current quoted turnaround is 21 days from receipt of samples. Questions about a specific result can be sent to admin@novagenlaboratories.com.

When a figure from any source needs checking, the first step is to confirm that the document is genuine. A Novagen certificate can be confirmed against the register at the verification page. For the next step in reading a result, see retention time explained.