Could this be due to an error in Mass Spec calibration or data analysis? I have 2 technical repeats that are fine, but the 3rd repeat is far away in the PCA plot and clusters with replicates of a different sample.
There are a lot of possibilities, most are related to the reproducibility of the sample preparation and reagents used therein. There can also be problems with calibration of the MS and changes in the elution times from the LC due to a buildup of material on the column. The problem with MS is the size of the feature space. Basic chemometrics 101 says that you need a minimum 3 independent samples for every feature to gain any statistical insight. When you have thousands of peaks and three samples, the probability of false positives by PCA is huge. No one has the money to run 3-5,000 independent samples.
Dear Joshua,
First, yes it can. As Luke V Schneider mentioned, there are many potential technical reasons to lead to this output.
But, I would first explore your data manually and carefully. The "outlier" on PCA plot can be a result of incorrect data processing that could be linked to a specific feature/some features that will dominate localization of the sample on PCA. So I would check the numeric data that you have used for PCA to check if this problematic sample is different from its replicates for all features, or just one/several specific. You can do it quickly by calculating CV between three replicates and see if it is very high for all feature, or just for some of them.
It is always better to include an internal standard (deuterated isotopes or any analog compound) to uncover any system or operator-related bias...
If you don't have any IS in your experiment, an alternative way to do this is tracking the intensity (and RT) of an inherently available compound (consider this approach as visualization of housekeeping proteins as normalization targets in western blot analysis).
If the system is Orbitrap, EASY-IC calibrant performance may also indicate some hints. A baffled system for ESI fluidics infusing the leucine enkephalin in the Waters system may also be beneficial...
shifted position for a sample in PCA is more common for biological replicates but not for technical replicates. If the instrument performance is stable during analysis, it is probably caused by either sample prep or autosample operation failure.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication