Hi there,
I am reprocessing data (PCA) I processed in metaboanlyst in 2021, with the same input data I get very different molecules high in my PC1 loadings, specifically the molecules we then did targeted analysis of, due to 7 of them being in the top 25 of PC1 back then, 6 of them are no longer even in the top 100. Is there any way to work out why this is and to find another way to replicate the results?
The most likely reason is that you used different normalization method(s) which PCA is most sensitive to
Didn’t do any normalising other than glog in v4 then log_10 for v6. Any other ideas?
Read and follow our post guideline