Skip to main content
Topic: File conversion for XCMS (Read 2251 times) previous topic - next topic

File conversion for XCMS

Hi guys,

I'm sure this is a fairly simple problem, but being new to XCMS it's a tad of a mystery to me.
I've been converting raw data into CDF files using Databridge.
The data transformation using DataBridge creates 3 different CDF files from a single original raw data set. Should that be the case?

If I load these in R they really clog up the system...they are huge. Do I need all three of them?

Cheers and much appreciated!

Re: File conversion for XCMS

Reply #1
You get one file per "function" with databridge. So you need to know how your experiment was set up to know which to use. Typically you might have a normal MS1 function and the lockmass function (not sure the latter is written as a file but might be). Then you might have added an MSe function. You can check what each are in masslynx.

What do you mean by clog up the system? XCMS doesn't load all raw data at the same time.

Re: File conversion for XCMS

Reply #2
Hi Jan,

Firstly, thanks for the reply.

I've been given the data so don't have much information on the experiment on hand.

My problem is that when I use Databridge to convert the raw file, it spits out 3 CDF files.
I have around 100 raw files. If I then try to load the newly converted CDFs into R using XCMS, the R or probably more likely the computer cannot (understandably) handle the huge size of these files.

Uploading to CDF is just as cumbersome as it takes forever.

Is there another way to convert these and make them smaller for use with XCMS in R or online.


Re: File conversion for XCMS

Reply #3
You need to figure out which of the 3 files you need. Mixing different functions will likely mess things up.
Since you have Databridge I guess you have masslynx. So open a chromatogram --> display --> TIC. Here you have the functions listed and you can get the TIC of each function.
If that doesn't clear it up either ask the people that did the experiment or try to decipher the _extern.inf in the raw folders. That contains all the experiment settings and have sections for each function. The format is not very consistent between versions but I just checked and it seems MSe functions will have things like
Transfer MS Collision Energy Low (eV)      10.0
Transfer MS Collision Energy High (eV)      40.0
listed at least in my files.

What do you mean by "load the files with XCMS"? Using xcmsSet? How heavy this is depends on the settings. What happens? You run out of memory? At which point? How large are the files? Maybe they are not centroided?


Re: File conversion for XCMS

Reply #4
Thank you Jan!

I'm completely new to MS, having done genomics before so getting going is a bit of a learning curve right now.

Just checked the TIC on a few of the RAW files, the three functions are all the same? TOF MS (50:1200) ES+

To the last part, yes, using
> list.files()
> xset <- xcmsset()

The files I am looking at are around 1GB each (that times 3 for each Raw file, but as you say I should only use 1 of the 3). If I only used 1 of the 3 that would bring down the total to around  40GB

Re: File conversion for XCMS

Reply #5
When you have files that are 1GB they are almost certainly in continuum mode.You need to first convert them to centroid mode in masslynx to be able to use XCMS. Typically centroid mode files are 50-100MB.
Masslynx --> tools --> accurate mass measure --> Automatic peak detection.
Then convert the resulting raw files.

As for the functions I am afraid also MSe is listed as TOF MS. So you probably have to ask the people that did the experiment what each are unless you can guess from that  _extern.inf file. But that probably requires comparison with something you know what is unless you are a really hardcore MS person.

For some info on XCMS I can plug my own tutorial here:

Re: File conversion for XCMS

Reply #6
Thank you!

Sorry, but is there an alternative to MassLynx. It just keeps hanging too much...

Re: File conversion for XCMS

Reply #7
To view raw files? No. To browse converted files you can use mzMine.

To do the centroiding? Yes, msconvert from Proteowizard can, but from the docs seems not well. msconvert cannot use the Waters' centroiding as it can for other vendor formats. So it has to use its own supposedly inferior implementation.

Re: File conversion for XCMS

Reply #8
Thank you!