Usage of MSnbase objects in xcms3 about xcms HOT 4 CLOSED

OnDiskMSnExp: this object might replace the xcmsRaw object. It combines all raw data from a whole MS experiment (reading the data on-demand from the disk). The data itself is organized in spectra, one spectrum for each scan/retention time. Extraction of spectra data is thus straight forward. Extraction of chromatogram data might be more complicated (time consuming).

Some notes:

• Use a as.matrix or as.data.frame method on the Spectrum1 classes to get the M/Z - intensity value pairs from each scan in one call, cbind that data and pass it to the low level do_ low level methods.
• getEIC or similar methods: is there something available for MSnExp/OnDiskMSnExp?

from xcms.

just want to chime in on this…

Over the years xcms has had many functions removed from it sometimes even becoming whole packages themselves, mzR is one example. I would feel that unless OnDiskMSnExp from msnBase has huge benefits over xcmsRaw in terms of reading data faster, compressing on storage etc that we stick with xcmsRaw. Many developer use this structure in their packages both public and private. I think we want to avoid xcms becoming a wrapper of mzR and msnBase.

Also happy to lend a hand in xcms3 re-write if I can be of any help, just let me know.

On Jul 27, 2016, at 11:31 AM, Johannes Rainer [email protected] wrote:

OnDiskMSnExp: this object might replace the xcmsRaw object. It combines all raw data from a whole MS experiment (reading the data on-demand from the disk). The data itself is organized in spectra, one spectrum for each scan/retention time. Extraction of spectra data is thus straight forward. Extraction of chromatogram data might be more complicated (time consuming).

Some notes:

Use a as.matrix or as.data.frame method on the Spectrum1 classes to get the M/Z - intensity value pairs from each scan in one call, cbind that data and pass it to the low level do_ low level methods.
getEIC or similar methods: is there something available for MSnExp/OnDiskMSnExp?
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub #41 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/ACJZDNL-K0WLYsgGFbU9uEJMHxQAyUDlks5qZ6QWgaJpZM4JWfEu.

from xcms.

Dear Paul I really appreciate discussions on this! There are no plans to remove xcmsRaw immediately.

For me the obvious advantage of OnDiskMSnExp would be in the following use-case:

• Features are identified (we have an xcmsSet) and we want to get the EIC for these across multiple samples. Currently we would call e.g. getEIC, which will then however read the whole raw data from each file. With the OnDiskMSnExp on the other side, only the required data will be extracted from the raw files. So there will be a huge performance gain here. Also, the OnDiskMSnExp class keeps track of any data manipulations that have been performed (also such as baseline correction etc).

from xcms.

Using the OnDiskMSnExp object in xcms as a container for the raw data and the XCMSnExp (defined inxcms and extending the OnDiskMSnExp) has proven to be useful. Biggest advantages are the direct link to the raw data, all already implemented and reusable subsetting and filter methods.

from xcms.