Peaking through fingerprints: Classifying medicinal herbs with detailed chemical fingerprints

Skip to Navigation


  • Published: Nov 7, 2011
  • Channels: Laboratory Informatics
thumbnail image: Peaking through fingerprints: Classifying medicinal herbs with detailed chemical fingerprints

Peak patterns

Chemical fingerprinting is a useful and effective method for classifying materials according to their chemical composition. The idea is that the pattern of peaks produced when a material is analysed by some from of chromatography can be used as a unique 'fingerprint', allowing the particular material to be distinguished from other materials. More often than not, this is done by using multivariate statistics to create classification models from the fingerprints.

Understandably, then, the more peaks that make up the fingerprint, the more specific the fingerprint will be for the material and the less chance of an incorrect classification. This can be important for material from living organisms such as plants, as the chemical composition of the same plant species can differ quite a bit according to factors such as climate, growing location and harvesting time. If you have a chemical fingerprint with lots of characteristic peaks, however, then the chemical similarities between different individuals from the same plant species should outweigh any differences.

On the right wavelength

To produce fingerprints with lots of peaks, three chemists from the Universitat Autònoma de Barcelona (UAB) in Spain have now come up with a rather novel approach. When producing a fingerprint of a material, scientists usually simply generate a single chromatogram by monitoring the absorption of ultra-violet (UV) light at a single wavelength. This is quick and easy, but obviously only detects analytes that can absorb UV light at that particular wavelength. In many cases, this still produces a fingerprint with sufficient complexity to be able to classify materials correctly, but not always if the materials are of biological origin.

So the UAB chemists, led by Jordi Coello, instead monitored a wide range of different wavelengths, allowing them to detect a lot more analytes. They then produced a fingerprint from just those wavelengths that detected the greatest number of analytes.

They first tried out this method on the roots of the medicinal herb Valeriana officinalis, which is used in traditional medicine as a sedative and to treat insomnia. V. officinalis can be sold in tablet form and some unscrupulous producers are not above adulterating these tablets with other, cheaper members of the Valerianaceae family, which are not known to have any beneficial pharmacological effects. Detecting this adulteration can be difficult, though, because it requires distinguishing between the natural chemical variation in V. officinalis and the slightly larger variation between different members of the Valerianaceae family. Hence the need for both fingerprints with lots of peaks and multivariate statistics.

Model building

Coello and his colleagues analysed various different samples of V. officinalis root using high-performance liquid chromatography with a diode array detector, allowing them to monitor over 150 different wavelengths. Analysing the chromatograms generated at these wavelengths revealed that the majority of the chemical information was present in the chromatograms produced at just four UV wavelengths.

After removing unwanted background noise and any unavoidable variation between chromatographic runs, the chemists merged these four chromatograms into a single version containing all the chemical information. They did this for each of the different V. officinalis samples, producing detailed fingerprints for all of them.

Then, by applying either soft independent modelling of class analogy (SIMCA) or partial least squares discriminant analysis (PLS-DA) to the fingerprints, Coello and his colleagues built two separate classification models. Not only could both models accurately distinguish V. officinalis from other members of the Valerianaceae family, but they could also distinguish V. officinalis samples that had been adulterated with various concentrations of other members of the Valerianaceae family. They could do this even if the adulterant was only present at concentrations of 5%.

Following this success, Coelo and his team are now looking to apply their novel fingerprint method to other medicinal herbs.

The views represented in this article are solely those of the author and do not necessarily represent those of John Wiley and Sons, Ltd.

Peaking through fingerprints: Classifying medicinal herbs with detailed chemical fingerprints

Social Links

Share This Links

Bookmark and Share


Suppliers Selection
Societies Selection

Banner Ad

Click here to see
all job opportunities

Most Viewed

Copyright Information

Interested in spectroscopy? Visit our sister site

Copyright © 2018 John Wiley & Sons, Inc. All Rights Reserved