DxOMark tests and scores camera image sensors separately. To test a lens you have to have some kind of image sensor. DxOMark choose to use image sensors for which the lens is intended. As a result they tested Nikon lenses on a very large number of Nikon cameras; an ambitious goal, but unfortunately slow and expensive. This might very well have been the beginning of their downfall. Trying to sell lens correction software, post processing software competing with Adobe, and finally their own camera does not seem to have helped.
The problems with testing lenses with a laboratory quality image sensor are two fold:
1. If you replace the sensor you loose the apples to apples comparison with earlier lenses tested.
2. You still do not know exactly how the lens will perform on a specific camera.
However, since the resolution of camera plus lens is the product (multiplication) of the transfer functions of the two, you can still produce a decent estimate of the image quality.
Anyway, please Nikon, when I buy my next Nikon lens I would like to see a performance specification showing center resolution, DX side resolution and FX side resolution expressed in LP/mm (line pairs per mm) as a function of aperture opening.
Thank You Nikon !
Truly Yours.