I feel I must chime in here and sound a small alarm bell. Please, please can we stop using the DxO ratings to compare between different camera systems? Madness lies in those charts, madness, I tell you!
First of all, nobody really knows how the final algorithm for the DxO rating is weighted or derived. So, from a simple First Rules basis - don't make decisions based on statistics that we don't understand. And I don't think anyone but DxO knows how they come up with their final rankings.
Secondly, the DxO scores may be fine to compare within brands - use it to look at the various Nikon cams, for example, but I think one can not rely on DxO scores to tell you much about how different makes of cameras compare. And here is one reason why: DxO measures noise, among other things, and these noise stats go into their final rankings. But Nikon sensors process out chrominance noise on-chip before the RAW image is outputted - other makers do not do this. This is an inherent bias, and there are most likely others. DxO rewards Nikon's strategy in their rankings, but this does not tell you what we need to know, which leads to...
Thirdly, the DxO rankings do not tell us which sensors/cameras can produce the best images.What they tell us is only which cameras produce the best DxO scores. In fact, Michael actually pointed this out in a tangential way when he spoke of the remarkable prowess of the Canon G10 compared to a medium-format digital back under certain conditions. If one was to look at the DxO chart, the G10 has one of the lowest overall scores on the chart, yet its IQ, under the right conditions, is comparable to the the highest scoring camera. This alone tells us that there is a huge disconnect between a Dxo score and the actual IQ potential of a camera.
The Nikon cams generally test relatively higher than the Canons because of their preRAW processing, but this does not tell us whether a Canon 5D or a D700, or a D3x vs a Sony 900 will produce the best images after optimal PP, which is how we use these systems, after all. And speaking of actual use, each of us has different ecological niche for our cameras. If I had a Canon G10, I would use it for landscapes without huge dynamic range requirements. That is, in the cameras sweet spot. We all would do this for any and all of the cameras we use. And this is natural, expected, and completely screws up the statistical significance of the DxO rankings.
So, I don't think we can rely on the DxO numbers to simplify our decisions as tempting as that may seem. We still need to test the actual cameras in the field, process the images, and make decisions based on a more challenging dimensional map than the one provided by DxO.