Chromatography Forum

Hello everybody,

I'm trying to use AMDIS to get area data for several hundred GC-MS runs.

As promising as it looks, I face a problem of non-identification. Peaks are identified as components, fall within the RT window and clearly have the same MS spectra of known compounds in the target library, but are not identified as targets.

In fact, with one sample I couldn't even get a "circular identification": I added the component to the library and reran the analysis, but the peak was still considered an unidentified component rather than a target compound.

Anybody has experience with this issue?

I have experienced something similar (and several other interesting idiosyncracies). Try changing the width parameter, the peak subtraction setting, and the match parameter.

If your name is a clue to what you are analysing then many of your spectra will have major fragments in common. Try the exclude m/z setting to make the deconvolution ignore fragments that are common to most of the compounds. I am doing more diverse mixtures, so I have not tried this.

Peter

Hi Peter,

Thanks for the quick reply!

I found that having a more inclusive minimum match factor (70) gives good results. I work in a "closed environment" as I build my own library based on compounds I identified with Chemstation, so there are not too many false positives.

As you mentioned, mono- and sesquiterpenes tend to confuse AMDIS a bit, but since it's not too many compounds I'll probably just correct the results manually.