What role does clustering play in the univariate ROC analysis?

MetaboAnalyst
1

Hello Omics Community,

I’m exploring the biomarker-analysis process of MetaboAnalystR and have come across a detail I cant really understand:
When I apply the univariate ROC analysis with MetaboAnalystR I get a table with the respective AUC, Pval and FC. for the respective compound. In addition there is also the column “clusters” indicating a cluster-number for each compound. What role does clustering play in the univariate ROC analysis?

Many thanks for your help and best regards,
André

2

Our tool is primarily designed for developing multi-biomarker models. The cluster number indicates similarities of different biomarkers (whether they are in the same clusters). You can use this information when you manually build a multi-biomarker model. In general, you choose biomarkers that are complementary to each other (i.e. from different clusters).

3

Thank you for your quick response! Very helpful.

4

This topic was automatically closed after 4 days. New replies are no longer allowed.