Can OPLS-DA Be Used If R² Is 0.93 but the Permutation Test Slope Is Negative
In OPLS-DA, the R² value quantifies the proportion of variance in the data that is explained by the model. An R² value of 0.93 indicates that the model accounts for 93% of the data variability, which is generally considered high and suggests strong explanatory power.
During a permutation test, a plot is typically used to illustrate the relationship between model performance metrics (such as R² or Q²) and the permutation order. In this context, a negative slope may suggest that as the permutations increase—i.e., as class labels are increasingly randomized—the model’s performance deteriorates.
If this pattern is observed, a negative slope is usually considered favorable, as it implies that the model’s high performance is unlikely to result from random noise. In a valid permutation test, a robust model should demonstrate high R² or Q² values when evaluated on the original (non-permuted) data, and substantially lower performance when the data labels are randomly permuted. If your model exhibits this behavior and achieves an R² value of 0.93, this generally supports the reliability of the model.
MtoZ Biolabs, an integrated chromatography and mass spectrometry (MS) services provider.
Related Services
How to order?
