Skip to main content ↑ ↓
Topic: Creating a msp library of pseudospectra from GC-MS data (Read 5890 times) previous topic - next topic

Creating a msp library of pseudospectra from GC-MS data

Hi XCMS experts,
I am a newbie in xcms and R. I am trying to analyze GC-MS metabolomic data (single quad) using xcms. I have one group of “Treated samples” and another group of “Control samples” to compare.
Using xcms, I managed to perform “peak identification”, “1st peak matching”, “RT correction” “2nd peak mathing”, “PeakFillIn” and “AnalysisVisualize”. By using “diffreport” function, I was able to create a *.tsv file with 3440 rows showing all m/z peaks identified at all retention times among the different samples. I can now determine the ions/peaks that are highly differentially expressed (using data from the “fold” column).
By sorting the data in ascending retention time, I noticed that there are parts of this dataset showing several m/z ions with similar retention times and similar fold-change. Since, these ions are obviously GC-MS fragments of the same compound, l want to extract pseudospectra for compound identification.
By using CAMERA, I performed xsAnnotate and groupFWHM. That gave me 569 pseudospectra, which after groupCorr resulted in 1124 groups. I was also able to write each pseudospectrum in a separate csv file using getpspectra/write functions inside a loop. I was glas to see that the pseudospectrum of my Internal Standard was reliably identified though a NIST library search... but checking all other pseudospectra manually will be time-consuming.
I would like to ask you if there is any easy way to create a msp library of pseudospectra that I could then easily search (as a batch) using NIST MS Search software.
Thanks
Manolis