Retention time cropping

[Yeshurun_Levi]

Hello everyone,

I am performing untargeted LC–MS analysis of samples containing a wide range of metabolites. My understanding is that peaks appearing at high retention times (RT) can sometimes result from column cleaning or late-eluting background compounds and therefore may not be suitable for downstream analysis.

As shown in the figure I shared, which displays pooled QC samples (mixtures of all samples, injected repeatedly throughout the sequence to mitigate run-order or batch effects), the data were acquired in alternating polarity mode. In the positive-ion data, there is a series of high-intensity peaks appearing toward the end of the chromatogram. These peaks show large gaps in RT between injections, despite having very similar m/z values. Notably, their m/z values are also close to those of distinct peaks observed earlier in the run.

Could this pattern reflect a bench or run-order effect, or is it more likely related to column cleaning or late-eluting artifacts? In either case, would it be advisable to exclude this late-RT region from the analysis?

Thank you

[xia.lab]

I see your point - better investigate these peaks w.r.t injection orders. QC is critical and you should try your best to address this drift issue as much as possible.

Since you are already at this forum, there are several options in MetaboAnalyst that might be helpful:

1. Batch correction with QC (the “utility”)
2. QC based on filtering (i.e., based on RSD)
3. Visualization via PCA or heatmaps (i.e. are QCs clustered v.s. other samples)

They will give you some sense of the impact on the data analysis and interpretation. You can always filtering these peaks based on their RT (manually) and compare.