Skip to main content
Topic: Duplicate annotations from Compound Discoverer (Read 2758 times) previous topic - next topic

Duplicate annotations from Compound Discoverer

Hi all,

I'm seeing a noticeable number duplicate annotations for the analytes generated from Compound Discoverer. Specifically, there are cases where the molecular weight are the same, but some (or all) of the retention times are different.

What is the best way of handling such cases during downstream analyses?

Thanks!

Re: Duplicate annotations from Compound Discoverer

Reply #1
Hello djb17,

I am experiencing the same problem. Did You find a possible solution?

Thanks

Re: Duplicate annotations from Compound Discoverer

Reply #2
Hi djb17 and AMO,
This thread is already a bit older, but I thought an answer might still be helpful for others in the future.
Duplicates in CompoundDiscoverer are very common and originate from a multitude of reasons.

If duplicate compounds have near identical Mass and Retention Time, your Compound Grouping settings might be set too strictly. Chromatographic shifts can be corrected by alignment, but even after that, the same compound might shift a few seconds during your sequence. Hence, you need to allow CD enough room to group multiple detections. The same applies to mass accuracy and the resulting MW (mz). I cannot suggest default values since it depends on your method.

In addition, CD can have difficulties with adducts, finding e.g. the M+H and the M+NH4, then reporting both as separate compounds. I would suggest adding suggested/expected ions to the Detect Compounds node.

Once the "processing duplicates" are addressed, you get to the more complicated but essential issue of duplicate annotations.
You mentioned you see the "molecular weight are the same, but some (or all) of the retention times are different".
That indicates you are annotating based on exact mass. Many compounds, such as isomers, might have identical mass but different retention times. Significant changes (method dependent) in retention time indicate actual different compounds. CompoundDiscoverer(or any software for the matter) just cannot tell these compounds apart by mass alone.

Possible solutions to elucidate which peak is the "true" compound, are MSMS fragmentation spectra and reference standards, run on your instrument.

In general, it's important to double-guess the annotation and compounds CD provides to you. Emma Schymanski et al. summarised the annotation levels in  "Identifying Small Molecules via High Resolution Mass Spectrometry: Communicating Confidence" very well. Have a close look and ask yourself where each of the Compounds in your CD results stands on the scale.

Finally, even when you know more about the Compounds detected and their annotation, you should still filter through your CD results. Peak quality parameters, such as S/N, precision over multiple injections, annotation sources, MS match scores and gap fill status, give you numeric parameters that help you filter your results, either directly in CD or in software after export (R, Matlab, Excel).

In any case, good question and good luck with your analysis, where ever you already are at this time point. :)