Block.spls and indY

Analysis
#1

Hi all,

I am trying to find a really correlated fingerprint which cross-sectionally describes my population by using 3 omics. For this purpose I have selected block.spls method such is not a supervised analysis. As I do not have an outcome Y, the method suggests to supply an indY which “indicates the position of the matrix response in the list X”. Could someone bettwer explain me, maybe with an example, what does it mean? should I indicate the design matrix?

Hope someone could help me!!

Thank you so much!

Serena

#2

hi @Serena,

I agree this is still a bit unclear! In block.spls() you still need to input a response matrix Y (very similar to a normal PLS that does a multivariate regression). Y can be a continuous matrix. I suggest you set one of the omics as Y (e.g. transcriptomics) and the two others as the Xs.

Otherwise, consider using wrapper.sgcca. Here is a summary, all functions are in the package:

Screen Shot 2021-02-10 at 15.57.41
Screen Shot 2021-02-10 at 15.57.41837×604 70.3 KB

Kim-Anh

1 Like
#3

Hi Kim,

Thank you so much for you clarification, it’s very useful!! I was actually trying to consider the numerical matrix of IDs of subject as Y, but not sure if that would be the best approach. I will try wrapper.sgcca as you reccomended :wink:
Again many thanks for your reply!!

Greatings,
Serena