PLS-DA analysis:'X' and/or 'Y' must be a numeric matrix

Support
1

Greeting,
I am currently using PLS-DA for my classification and discrimination analysis on my differential gene expression analysis for developing maggots. I followed up instruction. Here are all the commands I used and what it showed:

X <- t(Trinity.gene.TMM.EXPR.wAnnot.matrix)

dim(X)

[1] 10 14

test1 <- c("100","100","100","110","110","110","120","120","120")

Y <- factor(test1)

>levels(Y)
[1] "100" "110" "120"

summary(Y)
100 110 120 
  3   3   3

My input file looks like this

Screen Shot 2022-07-05 at 6.39.38 PM
Screen Shot 2022-07-05 at 6.39.38 PM1379×521 129 KB

And I received this error when I did MyResult.splsda <- splsda(X, Y, keepX = c(50,50))

Error in Check.entry.pls(X, Y, ncomp, keepX, keepY, mode = mode, scale = scale, : 'X' and/or 'Y' must be a numeric matrix.
I have tried as.numeric(X) or as.numeric(X) . The same message still persists. My original input X file looks like this:

I am not sure if the format causes the issue, but if you have any suggestions, please let me know, thank you!

2

Looking at your code, I’m intrigued as to why you transposed the Trinity.gene.TMM.EXPR.wAnnot.matrix. Which of the dimensions that are printed via dim(X) correspond to the sample size? A potential cause of your issue is that the features are found as rows rather than columns.

I would however guess that the the presence of the X row (from the first image you posted) or V1 (from the second image you posted) is the problem. These are character strings and cannot be converted to a numeric nor can they be used by any mathematical method, including those found in mixOmics.