Choosing of method

I have a mIHC dataset that has information on the different cells present in each sample based on the biomarker association they show. this dataset has pre-treatment and post-treatment data with 2 panels in each having different sets of biomarker each panel (its not entirely different, some are common). Each panel has 2 type of data, being whole number of cells and relative number of cells in percentage. I am confused as to use PLS-DA on each of the panels or p integration on the 2 panels.


hi @DARKSage77

You would need to have either the same cells, (or the same samples) for N-integration, or the same biomarker panels for P-integration in order to integrate these data. From your brief discussion it does not seem to be the case.
A PLS-DA would be suitable to analysis each panel independently, pre vs post. Have a look also at the multilevel decomposition in mixOmics. Multilevel | mixOmics



Thank you soo much for your reply. I was able to continue with PLS-DA and sPLS-DA.