Normlaization problem

[luna maria Munoz]

  

Hi to all, 

I was hoping to get some help on the following problem concerning the R-script using Cardinal. In our lab we have been trying to normalize our data using the "normalize.reference" function. Using the following script we get the normalization we want but an error message : 

Unfortunately, we don't understand where the 0 values come from and if we try doing the peakbin afterwards we experience a problem : R-script crashes. 

We have had this problem in our lab for quite a while so if anyone can help with this normalization problem it would be greatly appreciated. 

Thank you, 

Luna Munoz 

[Melanie Föll]

Hi Luna,

it seems the normalization does only give a warning (or was there an additional error?). For each spectrum in which the feature 320 intensity is zero you will receive one line of warning saying that your reference = 0.

All intensities of these spectra will be set to zero because normalization is not possible. You can find these spectra with "which(iData(data_test[ ,320])==0)".

peakBin might crash afterwards because in spectra without any peaks it's impossible to integrate the area under the peaks from the mz list, so if you remove these empty spectra before peakBin it should work.

HTH,

Melanie

[Groupe Chaurand]

Hi Melanie, 

I still have a couple of issues with the normalization feature: before normalization my data is composed of the following :

 

I applied the command :

"which ( idata(data_test [320.2, ]==0)

If I use the normalization feature after. It runs successfully:

And I achieve the following images : 

                                                            Before normalization                                                                                                                                                                                              After normalization: 

                                                                                                                                     

There is a clear difference between the two images but I do not achieve the desired results. To my understanding, if I image the m/z value that I used for normalization I would obtain the following ratio : desired m/z / normalization m/z : since I am imaging my 320.2 peak with a normalization of 320.2 I should obtain a ratio equal to 1 and therefore my image should show pixels close to a 1 (linear scale). I can see this result if I apply a normalization on  Flex-imaging: 

No normalization : 

With normalization :

With the cardinal package normalization I see the opposite effect after normalization : all my pixels are less intense than before : very close to 0. 

Is it normal that my pixels don't have an intensity close to 1 when normalization (of m/z= normalization) is done ? 

Is there someplace I can find information on how the normalization function works ?

Thank you !!

Luna Munoz

Groupe Chaurand - Université de Montréal