Extract intensity from a predefined m/z list

[Marie Gorka]

Hello Cardinal users,

I am reaching you regarding an issue concerning peak picking.

To put the context : I have several MSI analysis (one per sample) over a year and in order to process my data for statistical analysis I need to create a data frame for each of my sample where the columns are the same m/z for all my samples.

The idea is therefore :

- to extract a global list of m/z taking into consideration all the m/z detected in all my samples and that combine the m/z that close (10ppm) . I already managed to do it with MALDIquant but I cannot pursue for what I want to do after. Can we do it too with Cardinal ?

- For each sample, create an intensity data frame where the lines correspond to the pixels of the sample and the columns are the different m/z from my combined reference list. 

Can anyone help me with these issues ? And how to process peak picking based on a reference list created upon several samples ?

Thanks a lot in advance !

Marie

[foellm...@gmail.com]

Hi Marie,

there are many ways to do this, it also depends on your data and aims.

The most common way is to combine your data first, then perform pre-processing on them together (no need for a common m/z list then), use mzBin() in case your files do not yet have the same mz axis, combine with cbind() and then pre-process the combined data.

I hope that helps,
Melanie

[Marie Gorka]

Hi Melanie.

Thanks a lot for taking the time to reply.

That's exactly the solution I have figured out to solve my problem.

However I know have another problem now regarding the RAM of the computer. Indeed with more than 30/40 small files the computer does not have enough Ram to process the calculations when combining the files. That is why I was wandering if there is any other option to solve this problem without combining the files.

If you have any other tips ?....

Thanks a lot.

Marie

[foellm...@gmail.com]

Hi Marie,

please make sure that you use Cardinal version 2.6 or higher and that your single files are not loaded into memory (imageData(mse) should return class: matter_matc). For further trouble shooting it would be helpful if you could post your R code and error message.

An alternative would be to pre-process each file individually, combine the mz values of all pre-processed files into one vector and bin this vector into a common mz list. This list of mz values can then be used with the peakBin() function to extract the original intensities for each peak from the raw profile data by using the common mz list as reference. Afterwards files should have the same mz axis but contain less information because all peaks are centroided and combining them should work.

Best,

Melanie