Hi,

I think that the area measurement makes sense, if there is no significant background, or if you can subtract that. In case of similar problems, I'd fit two peak functions (Gaussian, Lorentzian, Voight...) on the peaks, and the fit parameters will contain the change in width or shift (and also the area). To do the fit, I'd use Origin, LabVIEW, or Matlab, but of course there are many other appropriate tools.

Can you upload a sample dataset?

David

Dear Laura Paletto,

As explained above, the best starting point is use the ORIGIN software, which is freely available on the Net, and find the Full Width at Half Maximum (FWHM) using Lorentzian fit. If the two peaks are well separated, then you may fit the individual peaks separately, it is the easiest way. However, if the two peaks are near to each other, you can opt fitting two peaks. It is also very straightforward. Then you must know the instrumental width, which is normally quoted by the manufacturer. You can also measure it by recording the emission spectrum of nay source which give sharp lines. Secondly, also measure its line shape, whether it is Gaussian or Lorentzian. Normally the Instrumental width is a Gaussian.

Next you can use VOIGT fit in which you will have an option to fix the Gaussian width, which is the sum of Instrumental Width and Doppler width , and run the program to fit the profile. This fit will give you a precise value of the Lorentzian contribution.

For a simple analysis, just use Lorentzian single peak fit and compare the FWHM form the two sample. You will immediately find the difference.

with best wishes and if you need nay help, just send a message.

Aslam

Thanks for your answer. 

I am afraid I am not a physician but a biologist and I have never done all these fittings so I am not totally sure where I can start. If I can calculate a ratio as I explained above, how can I explain the differences? What does the ratio tell me?

Thanks for your suggestions

Laura

Hello Laura.

If we are talking about Raman spectroscopy it seems strange that the elastin spectrum presents only 2 peaks. Are you sure it is not dominated by fluorescence?

Raman peaks can be characterized by their area and more precisely by the area ratio with another peak used as an internal standard (as you proposed to do), but also by their position (related to chemical environment and conformation of the molecular bond), their width (related to the conformational freedom), their shape (presence of shoulders or subpeaks) which can be better seen using second derivative of the spectrum.

The software package you used to record these spectra should have tools to measure these parameters and calculate derivatives.

I suppose you measured the Raman spectra of tissues using a Raman microscope

Raman peak intensity is first related to the quality of the focus you achieved during measurement, to laser intensity and to system alignment. This is why Raman peak intensity is seldom used since all these can vary a lot during measurement.

The ratio of peak areas is a better choice (although crude) to extract information from the spectrum. It will inform on the relative concentrations of the molecular species that give rise to these peaks. If the 2 peaks come from the same molecular bond (in case of peak splitting) it will inform on the bond conformation or chemical environment (for example buried and exposed tyrosine).

 Laura:

It would be most informative if you could say what spectroscopy you are using. Raman spectroscopy (likely)? UV-VIS spectroscopy? IR/FTIR spectroscopy? If your sample is actually wet and possibly also below the surface of water, then I'd guess Raman. It's possible you don't even know, given your description. That's OK. Please find out and confirm.

As for analysis, the scientific graphing software called Origin, published by OriginLab Inc., is indeed very popular and very powerful. I use Origin in my lab. However, it is NOT free software as mentioned by others. If you're getting it freely on the internet, you're violating copyright law.

All the methods of spectral analysis described so far are valid. But you've mentioned that you are not familiar with these techniques. So let me suggest a couple of very simple methods that are easy to understand and implement.

First, you can look at simple spectral subtraction. Decide on a particular sample tissue that you can consider a kind of 'reference' sample. Record it's spectrum. I assume you have the data for the spectrum as a two column table (e.g., wavelength in one column and intensity in the next). Most spectrometer software allows you to 'export' or 'save as' a "CSV" file and that's what you want.

Once you have a reference spectrum, you can do 'spectral subtraction' with subsequent samples. That is, simply subtract the new sample from your reference. This difference spectrum will tell you something about how the new sample differs from the reference. You can use a spread sheet to do this (e.g., Microsoft Excel (not free) or Openoffice Calc (truly free)). Spectral subtraction cannot tell you everything about how the spectrum is changing, but it is very simple and it can reveal important differences. It's a popular technique in IR spectroscopy.

You could implement a curve fitting routine in your spreadsheet (via the 'goal seek' algorithm). But if you're not familiar with curve fitting (as you say), I don't recommend you try this. Even so, there are a number of tutorials on how to do this that can be found around the net. Perhaps a more advanced student in your lab can help you with this. Of course MATLAB, LabVIEW, MathCAD, etc. can all also do this. But again, you might not be familiar with these tools.

If you want to look at area, you can try a very simple 'physical' technique often used (in the past) by chromatography specialists. (OK, don't laugh now ... ) Carefully plot your reference and sample spectra at the same scale, but on different pieces of paper. Physically cut out the peaks of the spectra with scissors. Weigh the pieces of paper for each peak. The mass is an analog of area. You may have to approximate a hand drawn Gaussian peak fit. That's OK, your eye already has a good idea what a Gaussian shape looks like. Just approximate the peak. This sounds ridiculous, I know. But it's remarkably accurate as a rough approximation. If you insist on units of area, just measure the mass of a 1cm x 1cm piece of paper and scale the masses of your peaks accordingly.

As for measuring peak widths, red/blue shifts, etc. curve fitting is the best way to go of course. But that spectral difference method I mentioned previously can actually detect such differences. It's too complicated to explain fully here, but if you think carefully about what would happen to the difference spectrum if there was a red/blue shift, or if there were a change in width, you'll see how the method would reveal such changes. It's all relative to your reference sample, of course. But that might be all you require. If you want absolute numbers, however, curve fitting is the only way to go.

Laura, I hope you find this useful.

Good luck with your research and best regards,

Pat

Добрый день, Лаура! Я бы тоже хотела знать, какой конкретно спектроскопический метод Вы используете, В любом случае проводить на начальном этапе математическую процедуру разложения кривых вряд ли эффективно, поскольку у вас нет данных о параметрах спектров исходных форм реагентов и их изменений в ходе реакций эластанов с тканями. В любом случае требуется изучение растворов эластана в чистом виде и их изменение в зависимости от содержания в этих растворах компонентов изучаемой среды и возможно, других параметров, как, например, рН, температура и пр. Дальше, если нужно, можно обратиться к спектрам растворов с обеими тканями и, если нужно, приступить к процедуре разложения. Таким образом, на мой взгляд, должен выглядеть первоначальный этап эксперимента.

Пишите. Мой адрес: [email protected]

Успеха Желаю.

Софья Степанчикова.

Thank you so much, it's really helpful.

However, I haven't used Raman microscopy but I recorded the spectrum on a laser scanning microscope so maybe it won't be the same. I will try your suggestions for sure.

Thanks again, it was very informative.

Laura

Laura:

Well, it sounds like you're actually using a "Micro Raman" equipped microscope.These are quite common in bio research now-a-days. You might want to check the user manual to confirm exactly what kind of microscope you're using. Just remember, the more you know about the instrument you are using, the better science you will do.

Well, at any rate, best of luck with your research.

Pat