Re: peak separation
- To: mathgroup at smc.vnet.net
- Subject: [mg36132] Re: [mg36094] peak separation
- From: "Y.A.Tesiram" <yas at pcomm.hfi.unimelb.edu.au>
- Date: Fri, 23 Aug 2002 00:25:04 -0400 (EDT)
- Sender: owner-wri-mathgroup at wolfram.com
Hi Heinrich, There is a book available called Advanced Tutorials for the Biomedical Sciences: Animations, Simulations, and Calculations Using Mathematica and you can get a description of it via, http://store.wolfram.com/catalog/books/sci.html. However, there are several things that you can do yourself in Mathematica. Firstly, any equation that you find in a textbook such as an Analytical Chemistry or Instrumental Analysis book can be handled in mathematica with ease. o Taking the case of chromatography, if your data is in ASCII format, you can directly import this into Mathematica using Import or ReadList. Binary data can be handled as well and the package BinaryFiles loaded with, <<Utilites`BinaryFiles` will be your best friend in those cases. o Then I would plot it using ListPlot to see that what you have is what you intended to have. o Once you have this, your options are open. To separate overlapping peaks, some form of fitting will be required. In this case you will need the Statistics`NonlinearFit` package so you can fit your data points to a the gaussian function or indeed an function that resembles a peak (Lorentzian?) o However, because your peaks overlap you probably only have part of your points that belong to one peak. The rest are buried under the other peak. So I would use an interpolating function across the overlapping peaks, locate the maxima and fit the points that you can see from the located maxima to the theoretical form of your peak shape (gaussian etc). There may be some hiccups here, but with a little trial and error some good fitting can be obtained. o Of course you would have standard runs that would give you an idea of what the peak shape parameters should be. I would start fitting with these first. Hope this helps you get started. All the best Yas On Wed, 21 Aug 2002, Heinrich Kestler wrote: > Hello NG, > > anybody out there who ever tried to separate overlappeded peaks (eventually > for original signals of different types like gaussian, cauchy etc.), for > example im gas-chromatography or x-ray crystallography using > mathematica? > I would be greatful if I could receive a hint how to solve such kind of > task. > > Many thanks in advance, > Heinrich > > > > > > >