PLS-DA: Multilevel/group calculations (n>2)

MetaboAnalyst
1

Hello MetaboAnalyst team,

I am trying to recreate the PLS-DA calculations that MetaboAnalyst does with Python’s ScikitLearn PLSRegression function. In MetaboAnalyst’s Metabolomics Data Analysis powerpoints (in the Tutorials section), I get the impression that the PLS-DA calculations are only done on two factor levels/groups at a time? I also see in this PLS-DA for Python tutorial that “If you have more than 2 groups, but want to perform a similar 2-group analysis, you can filter the DataFrame using Pandas indexing.”

How does MetaboAnalyst handle more than 2 levels/groups in its PLS-DA calculations? Are the calculations applied across the whole dataset or is the data automatically subset in some way?

Any help would be much appreciated.

Kind regards,
Jessica

2

PLS-DA can handle multi-groups. The underlying MetaboAnalystR is in our GitHub. Specifically, the PLS-DA is a wrapper of the pls.da function from the caret package

3

Hi Jeff,

Thanks so much for providing those links, the information is very helpful!

Kind regards,
Jess