Understanding the correlation between components

[adhikaria]

I am performing N- integration in the RNAseq and proteomics data sets. While calculating the correlation between different components I got the following results:

Correlation between the first component: 0.99
Second component: 0.99
Third component: 0.98

Can anyone please help me understand what does such a high correlation between components means and how to interpret the results? Also in the figure, in the first components, the ellipses are not overlapping but overlapping in the second and third. Does it mean anything about the result?

[kimanh.lecao]

hi @adhikaria ,

This has been explained on our website and vignette. The method aims to maximise the correlation between components so it’s a good sign that your correlation is high (as a check that the method is working well / that your datasets contain similar patterns or information). Passed component 2, there is no discrimination between your two classes so your interpretation should focus mainly on component 1 (use the second components for 2D plots though).

Kim-Anh