I am currently using CAMERA function for metabolite annotation. I know that CAMERA is directly working on the XCMS object and do the annotation based on the peaks after XCMS fillingpeaks. By calculating correlations within one sample and accross samples for peak grouping. But this correlation is done based on the peak intensity. In our case, we know that we have signal intensity drift within one analytical batch(220 samples) and even more serious intensity drifting between batches (8 batches). So I am worried that the correlation function cannot work well to group peak together.
Do I misunderstand the CAMERA ? Any suggestions?
Thanks a lot in advance and I am eager to hear ideas from you.
The intensities of ions originating from the same molecule should "drift" in the same way and therefore the answer is no.
If you changed something within your analysis that can affect the ratio of fragments then the answer is yes. I have seem routine maintenance done by the vendor's technician change the fragment ratios (never figured what was changed) so avoid this kind of thing.
I have seen cases where adduct/polymer formation have non-linear behavior and in those cases the correlation can suffer.
I don't recall seeing it but I imagine detector saturation can also cause non-linear behavior.