Thank you for all your comments! I think it would be great to have the "paired refinement" implemented in future updates of refinement programs.
Coming back to my original questions: if I have overestimated the high resolution cutoff, what is the correct procedure to refine against lower resolution data? Reprocessing the data from images to a lower resolution(xds..), truncating at the scaling step (aimless...) or setting the cutoff during refinement (phenix...)? How do I have to treat Rfree flags? Thanks!
