Error when using plotIndiv.sgcca

[Lluís Revilla]

I am trying to use plotIndiv on a result obtained from RGCCA::sgcca, however I am getting an error. I tracked the issue and it seems to come from the line 143 of the plotIndiv.sgcca.R file where object$X is called. In my object I don't have an X element, instead I have element Y. Changing this to object$Y would correct this issue.

[Kim-Anh Le Cao]

Hi Lluis, I believe you are trying to use different packages. plotIndiv is part of mixOmics only, not RGCCA. You could try use mixOmics::wrapper.sgcca and plotIndiv (which is part of mixOmics):

> ?wrapper.sgcca
> data(nutrimouse)
> # need to unmap the Y factor diet if you pretend this is not a classification pb.
> # see also the function block.splsda for discriminant analysis  where you dont
> # need to unmap Y.
> Y = unmap(nutrimouse$diet)
> data = list(gene = nutrimouse$gene, lipid = nutrimouse$lipid, Y = Y)
> # with this design, gene expression and lipids are connected to the diet factor
> # design = matrix(c(0,0,1,
> #                   0,0,1,
> #                   1,1,0), ncol = 3, nrow = 3, byrow = TRUE)
> 
> # with this design, gene expression and lipids are connected to the diet factor
> # and gene expression and lipids are also connected
> design = matrix(c(0,1,1,
+                   1,0,1,
+                   1,1,0), ncol = 3, nrow = 3, byrow = TRUE)
> 
> #note: the penalty parameters will need to be tuned
> wrap.result.sgcca = wrapper.sgcca(X = data, design = design, penalty = c(.3,.5, 1),
+                                   ncomp = 2,
+                                   scheme = "centroid")
> plotIndiv(wrap.result.sgcca)

[Lluís Revilla]

Precisely I thought that plotIndiv having a method for class sgcca would work with the output of sgcca of the RGCCA package. Sincerely having a "wrapper.sgcca" that outputs something of a different definition of the original object that it is mimicking is confusing.

[Kim-Anh Le Cao]

The wrapper.sgcca is a wrapper of sgcca. The results would be the same as RGCCA:sgcca if you used the same input parameters, but RGCCA does not provide any visualisation, so we had to tweak the function,

[Lluís Revilla]

No, the code of wrapper.sgcca doesn't call RGCCA (so it isn't a wrapper although does the same as far as I know) and the output is different from that of the RGCCA function. With your example and using RGCCA::sgcca:

> names(wrap.result.sgcca)
 [1] "call"               "X"                  "variates"          
 [4] "loadings"           "loadings.star"      "design"            
 [7] "penalty"            "scheme"             "ncomp"             
[10] "crit"               "AVE"                "names"             
[13] "init"               "tol"                "iter"              
[16] "max.iter"           "nzv"                "scale"             
[19] "design"             "scheme"             "explained_variance"

> result.sgcca = sgcca(data, C = design, c1 = c(.3,.5, 1), ncomp = c(2, 2, 2), scheme = "centroid")
> names(result.sgcca)
[1] "Y"      "a"      "astar"  "C"      "c1"     "scheme" "ncomp"  "crit"  
[9] "AVE"

As you can see one outputs Y and the other X, which makes the function of plotIndiv to fail as reported.