PLSDA On omics data

I am attempting to classify my data, but I am consistently experiencing poor separation. To date, I have employed PCA, t-SNE, sPCA, and PLS-DA techniques, yet the results remain unsatisfactory. My objective is to achieve better separation between the two groups: controls and cases. Notably, I have 14 control samples and 38 cases, with 1677 proteins serving as variables. And for normalization, I have applied scale=True .

hi @Oumaaa12,

Then it means your data are not separable.