As I've just been using a small number of light frames I load the raw images select 100% stacking so all images are used.
Other settings are 4% on star detection and a median filter before registering for reduced noise and then median combination method and RGB background calibration also.
Result mode std
Light mode median
Alignment automatic
Intermediate files are created for each light frame
and a debayered image saved when raw img processing.
Autosave an output tiff in 16 bit file which is reloaded to allow curves adjustment for contrast and saturation although I have not been adding any saturation really just improving the black point for better contrast.
I then save the final file to a new location that is then loaded into CS5 for colour adjustment in selective colour as the file looks almost black and white from DSS but the colours are soon brought out in photoshop.
That comes down to your personal choice then, as you are bringing the colour up from almost nothing you can easily over or under do the colours so you realy need to look at what is an accepted image by others I guess. Can be a bit of a fiddle to get right
