I’ve performed multiomics integration using OmicsAnalyst. I’ve choosen DIABLO as dimesionality reduction method and I’m able to visualize the loading score of each feature on each component from the 3D scatter plot interface. I would like to ask how can I now select the main important features, able to discriminate between my two experimental groups? I can do it based only on ranked absolute value of the loadings for each component? There is a way to perform a VIP analysis?
Hello Chiara,
As of now, you can do it based on the absolute value of the loadings for each component. Have a look at the scatter diagram and determine which component(s) separate most the two experimental groups and identify the top features in terms of loading values. I am considering adding VIP analysis in the future.
Thanks,
Guangyan