search

sneumann / xcms

This is the git repository matching the Bioconductor package xcms: LC/MS and GC/MS Data Analysis
Other
178 stars 81 forks source link

fillpeaks on data matrix #580

Closed ccshao closed 2 years ago

ccshao commented 2 years ago

Hello, I would like to know how to fill the "0" in the lipidomimcs data, with only a data matrix.

Briefly, we got a data matrix from collaborators from our LC-MS/MS experiments. It is a compound id by sample matrix, while the compound ids are in the format of "retentiontime_mz", like "10_1000m/z". The data is generated by waters progenesis. We know the groups of samples.

In the matrix there are a few "0", we interpret them as missing value but no really "0". Our collaborator suggest to try the fillpeaks in xcms.

My question is, is it possible to do so with only a matrix, and how? What else info do I need to generate the xcmsSet in addition to the matrix? Sorry for the quite basic question.

jorainer commented 2 years ago

The fillChromPeaks only works on an XCMSnExp which represents the raw data (i.e. data in the original mzML files) of an experiment as the algorithm fills in raw signal from these files. Without that you would need to impute missing values. There are various approaches to do that, but they are not implemented in xcms.

ccshao commented 2 years ago

Thanks for the update! Do you have some ideas on which imputation methods to choose? Actually I saw a few papers on this topic but am not sure how broadly their conclusions in those papers hold.

sneumann commented 2 years ago

Your best bet will be to test a few on your data for the kind of analysis you aim for. Yours, Steffen

ccshao commented 2 years ago

Thanks, I might try a few used for microarray studies.