The image above is a static capture of a dynamic OpenGL project I created in which I wrapped a NASA albedo, i.e., sans clouds, image on an OpenGL generated sphere. In so doing, I also generated the UV coordinates associated with each vertex position.
This was an incremental learning effort in which I had already applied model matrix corrections to the vertex positions for the rotating and translating (orbiting) "Earth".
I was surprised to find that I did not have to apply a model matrix correction to the UV coordinates.
I have tentatively concluded that once the jpg image coordinates are associated with the corresponding vertex positions with the UV coordinates in the range of [0, 1], they are fixed and need no further correction.
Does that sound correct, or is there more to the situation?