LC-MS: Can a "bad" peak still be interpreted

Question

I have done an untargeted metabolomics study using LC-MS. There is one peak that matches to a biologically relevant compound but analyte eluted at the beginning of the LC run and the peak shape is far from ideal.

Do I assume the peak is an artifact, or can I say that there is a compound with that m/z in my sample? Ideally I would run the experiment again with standards, but I cannot do that due to time constraints. Someone else will follow up on the interesting compounds that I've identified and I wouldn't want to send them in the wrong direction.

Below is the peak in question, and following it is an example of a "good" peak from the same dataset.

Answer 1

This is low-quality data. It's very difficult to say whether this signal is meaningful for several reasons.

1. It appears as soon as the LC-MS starts recording data at very low retention times. That means it could be an artifact of column regeneration. There is no sign that the data contains a "void volume" peak, meaning it is unclear whether the low-RT signal is from sample injection, from the previous sample, or some other artifact.

2. You don't say whether you randomized sample injection order. Therefore the differences between the red and black sets of samples could arise because of changing conditions in the LC-MS instrument over time (e.g. column fouling, sample aging in the autosampler) or from legitimate differences in the samples that are biologically meaningful.

3. You don't show the full mass spectrum for your proposed analyte, so it is possible that the single m/z you showed here has an isotopic pattern consistent with your analyte, or inconsistent with it.

Thus, to answer your questions:

1. It basically wrong on the basis of info presented to claim that you detected a "biologically interesting" metabolite on the basis of the information presented.

2. It is also bordering on misleading to say you detected "a molecule" with the given m/z at all. Maybe you could say you detected an unknown signal at m/z of xyz and are unsure of its origin. That's about it.