A chromatogram is the graph obtained at the end of a chromatography run, showing the detector signal against time. At first glance it looks like a series of mountain silhouettes. Read correctly, however, it tells you how many components in the sample have been separated, their proportions relative to one another and how reliable the separation is. The operating principle of the method was explained in the article on HPLC; this article focuses on how the resulting graph is interpreted.
Axes and basic concepts
The horizontal axis shows time, usually in minutes. The vertical axis is the detector response; in ultraviolet detectors, this response is expressed in absorbance units or in milli-absorbance units (mAU), one thousandth of an absorbance unit. The moment the sample is injected onto the column is the zero point of the graph.
The time taken by a molecule that does not interact with the column at all to reach the detector is called the dead time. The small disturbance seen at the start of the graph usually marks this moment. No component can elute before the dead time; the peaks that follow immediately after it are hydrophilic components that have been retained on the column hardly at all.
Many software packages also draw the gradient program as a second line over the chromatogram. Seeing at what proportion of organic solvent a peak elutes gives a direct indication of the component's hydrophobicity. However, the composition set by the pump reaches the column with a certain delay because of the tubing and mixer volume in the system; this delay is taken into account when chromatograms from two different instruments are compared.
| Concept | How to read it | What it tells you |
|---|---|---|
| Retention time | The time from injection to the highest point of the peak | A reproducible identity clue under the same conditions |
| Retention factor | The difference between retention time and dead time, divided by the dead time | How strongly the component is retained on the column |
| Peak area | The area under the peak curve | A signal proportional to the amount of substance reaching the detector |
| Peak height | The distance from the baseline to the peak apex | A quick estimate; less reliable for broad peaks |
| Solubility | The difference in time between two adjacent peaks divided by the average of their widths | How well two components are separated from each other |
| Symmetry | Comparison of the front and back halves of the peak | Tailing or fronting |
Peak, area and height
An ideal peak resembles a symmetrical bell curve. The peak area is related to the amount of substance passing through the detector and is the quantity used as the basis for quantitative assessment. Height carries similar information, but as a peak broadens, the same amount of substance gives a flatter peak; height is therefore regarded as a shortcut that works for narrow, well-shaped peaks.
Calculating the area is called integration and is done by the software. Using its own parameters, the software decides where the baseline starts and ends and where the boundaries of the peak lie; when the parameters change, different area values can be obtained from the same chromatogram. Dropping a perpendicular between merged peaks or separating a small peak from the tail of a large one with a tangent are two common choices that change the result. Stating manual corrections in the report is part of good practice.
Why is area percentage not mass percentage?
The area percentage often seen in reports is the ratio of each peak's area to the total area. This value should not be read as the mass ratio of the components, because the detector responds differently to each component. When measuring at 280 nanometres, a peptide rich in tryptophan and tyrosine gives a strong signal, a peptide with few of these residues appears weak, and a peptide with none of them is barely visible. The background to the choice of wavelength was explained in the article on 280 nm. Absolute amounts require calibration with a pure standard of the component concerned.
Quality of separation
If the signal between two adjacent peaks falls all the way to the baseline, the components are fully separated. This is generally considered to be achieved when the resolution value reaches about 1.5 or above. At lower values, peaks overlap and their areas affect each other.
Symmetry also tells you about the health of the separation. Tailing peaks, whose back side is drawn out, usually point to unwanted secondary interactions between the component and the column surface, or to column wear. Fronting peaks, on the other hand, are associated with overloading of the column or with the sample having been prepared in an unsuitable solvent. A shoulder on the flank of a peak is often the trace of a second component that has not been resolved.
In systems equipped with a diode array detector, spectra taken from the start to the end of a peak can be compared to gain further insight into whether the peak consists of a single component. This check is a useful clue, but it may not be enough to distinguish peptides whose spectra are very similar.
Reading step by step
- Read the method header: column, gradient program, wavelength and injection amount.
- Examine the blank run: peaks seen in a gradient run without sample come from the solvent or the system and are not attributed to the sample.
- Determine the dead time and exclude the signal before it from the assessment.
- Follow the baseline: slight drift over the gradient is normal; sudden jumps are not.
- Count the peaks; do not count fluctuations that do not clearly exceed the noise level as peaks.
- Look at the separation of adjacent peaks, their symmetry and any shoulders.
- Assess the area table together with the integration parameters.
- Overlay and compare with a reference chromatogram obtained by the same method.
A common rule for the noise threshold is to take the level at which the signal exceeds roughly three times the noise as the limit of detection, and the level at which it exceeds ten times the noise as the limit for reliable quantification.
Reading a venom profile
In a venom sample containing dozens of components, the chromatogram is read as a whole, like a fingerprint, rather than peak by peak. In our reference batch analysis, reversed-phase separation was carried out with a gradient in which the proportion of organic solvent rose from zero to sixty per cent, and 37 proteins and peptides were resolved at 280 nanometres. When two batches are compared, what matters is less the number of peaks than the retention times of the main peaks and their areas relative to one another; small shifts in an individual peak often remain within the limits of method variability.
A chromatogram separates components, but it does not name them. Which peptide a peak corresponds to is determined by complementary methods such as mass spectrometry. Our batch results and method conditions are published on the analysis page.
