Hi @Chiara.Anser,
Sorry for the confusion, you’re right we currently don’t have loo functionality for tune.spca() (we do for tuning other models hence my mistake). In this case given that you only have 6 samples I would extract the loadings of your variables rather than running sparse tuning to extract the key features, you can do this using $loadings on your PCA or sPCA object and you can also visualise these using the plotLoadings() function.
Cheers,
Eva