omicverse.metabol.s_plot

Contents

omicverse.metabol.s_plot#

omicverse.metabol.s_plot(result, adata, *, label_top_n=15, ax=None, figsize=(5.5, 4.5))[source]#

OPLS-DA S-plot: p(cov) vs p(corr), i.e. covariance vs correlation between each feature and the predictive component.

This is the classic visualization for interpreting OPLS-DA loadings (Wiklund et al 2008). Features in the two “arms” of the S (high-covariance, high-correlation) are the strongest drivers of the case/control separation.