#------------prepare loading data---------------------
plsda.loadings2$x<-“Element” #variables in my study are chemical elements, so variables group
plsda.loadings2[,“Deposit subtype”]<-“non” #this column is empty by I need it in order to have same columns that scores table
plsda.loadings2[,“Altered rocks”]<-"" #this column is empty by I need it in order to have same columns that scores table #------------prepare scores data---------------------
plsda.scores2$Type<-"" #observations column is empty because I label only variables
plsda.scores2$x<-“Sample” #group of samples
plsda.scores2<- mutate_if(plsda.scores2,is.numeric,~ . / 9) # scores values are modified in order to scale them to loadings
#-----merge scores and loadings--------------
biplot.data<-rbind(plsda.loadings2,plsda.scores2[,c(5,11,13,18:22)]) # merge
I hope that can help you.
Really this is not a biplot, it’s just a way to display scores and loadings in same plot.
Many thanks for the code. After install your “addon”, I am able to produce a biplot with PLS-DA.
However, I noticed that there were some other changes when running with mixOmics, for example, mypls$explained_variance was changed to mypls$prop_expl_var, and that created some issues my old code.
I also cannot run the perf function normally.
For the example:
X ← nutrimouse$gene
Y ← nutrimouse$lipid
MyResult.pls ← pls(X,Y, ncomp = 4)
perf.pls ← perf(MyResult.pls, validation = “loo”)
perf.pls$Q2.total now gives me NULL
Is there anyway that I can go back with the “original” function of mixOmics.
I managed to obtain a biplot for my PLSDA, but I am working with a huge number of variables, so the graph shows a cloud of arrows.
I would like to be able to select which variables are shown in the biplot, because I’m interested in some groups more than in others. Is this possible with the current version of the code?
I have tried sPLSDA in previous projects, but I am not looking for a way to select variables, maybe I didn’t explain myself very well.
I have a matrix of thousands of microbes, and I did PLSDA, and everything works fine, and I can create your biplots, but I can’t see anything interesting because I have too many variables, therefore, too many arrows in the plot.
What I want to do is to keep exactly the same PLSDA that includes thousands of microbes in the X matrix, but I want the plot to only show for example the methanogens. I know their names, and I wanted to be able to just feed the code with a vector of names, and have these appear in the biplot.