Hi all, I used spls for muli-omics analysis of a dataset with 5 data types and a continuous outcome. I have now data on one of these datatypes on another set of samples. Can I generate a score for each of these samples based on results of the performed spls? I was trying linear combination of loadings with data, but applying this method to the previous sample did not result in the same scores provided by the spls. Any clue to how the score is generated by spls? Is it feasible to derive scores for a new sample if one of the datasets is available using previous spls results?
1 Like
hi @zia,
Yes, you can use the function predict() (see the help file) and that will output the predicted components. You can then use the plotIndiv() function with the argument style = ‘graphics’ or 'lattice' and overlay this prediction with the training set using the base function from R text() .
Kim-Anh