Yes, autobuild will read the test set (freeR flags) and use these free R flags and write them out to overall_best_refine_data.mtz. So yes, the R-values are directly comparable if you refine against the input experimental data or overall_best_refine_data.mtz. As Luca mentioned, you could verify this with mtzdump or with:

From: [email protected] <[email protected]> on behalf of Sam Tang <[email protected]>
Sent: Wednesday, May 11, 2016 8:59 PM
To: PHENIX user mailing list
Subject: [phenixbb] FreeR flags from AutoBuild

Dear all

May I enquire a simple question about AutoBuild - I input an experimental data (ccp4 mtz from Aimless) and a model from Phaser-MR to AutoBuild hoping that some mis-fit flexible regions could be re-build. May I know if the output mtz overall_best_refine_data.mtz contains the same FreeR flags as in the input, or would AutoBuild generate a new set of FreeR? Or in other words, are the R values directly comparable if I refine against the input experimental data mtz or the overall_best_refine_data.mtz?

Thanks!

Cheers

Sam Tang

Biochemistry Programme, School of Life Sciences, CUHK