Integration parameters in masslynx and targetlyx

Hi.I am setting up a method and have been trying to find out about the following integration parameters; mean and SG.What is the difference between mean and SG? I was told it is better to use mean for spectrums and SG for chromatograms but why I do not know. I was also told to use what works best but I would still like to know more about them. If I was asked why did you pick mean over SG I would like to be able to give a reason other than it looked better!Can anyone answer this question for me?


  • Smoothing


    I have inserted some text from the MassLynx help files for reference:

    Two types of smoothing are available for chromatograms; Moving Mean and Savitzky Golay. Both methods slide a window  along the chromatogram, averaging the data points in the window to produce a  point in the smoothed spectrum. Moving Mean takes the arithmetical mean of the intensities of the data points in the window.  Savitzky Golay takes an average of the  intensities weighted by a quadratic curve. This tends to enhance peak and valley  shapes, as well as preserving the height of the peaks better than the Moving Mean. However, Savitzky Golay does tend to produce small  artifacts on either side of the real peaks.

    Once the parameters have been selected, click OK to return to the Integrate chromatogram dialog.

    I have attached a files showing a peak smoothed using mean (2x3) and SG (2x3)  The SG smoothed peak is a similar intensity to the unsmoothed peak but there are artefacts on the peak, this may cause some issues when integrating the peaks.  The mean smooth is a wider peak but it will provide a much more consistent integration. I have chosed parameters that illustrate the point.   As a rule of thumb, I use a mean smooth for chromatography choosing parameters that do not distort the peak.  In this case I would use a 2x1 mean smooth.


    Jody Dunstan.


  • Thanks Jody,

    I have asked so many people this question and no one seemed to know the answer. Much appreciate your reply.

Sign In or Register to comment.