Hello,
I have a statistical question. I have three biological replicates and for each biological replicates two injections (2 analytical replicates). When performing an analysis, I see no way to differentiate between the two types of data sets. I have concerns that with an n of 6 compared to 3, some intensities may be made more significant than they are. However, I clearly want to control for machine produced error. Does anyone have any suggestions for dealing with this problem?
Thank you,
Trammells
No, there is no currently no built-in method to differentiate between technical and biological replicates.
However, the biological variability should be much higher than the technical variability (you would have a problem otherwise). There are several papers available on that subject.
Therefore, having biological replicates is much more important than technical replicates.
Usually running at least 3-5 biological replicates is recommended for samples obtained under controlled conditions, even more if conditions cannot be controlled.
If you are concerned about false positive differential features then you might try running XCMS with and without these additional technical replicates to see if it makes a difference.
Ralf