Hello Jorge,
Kay has also responded and given some information as to how to potentially do what you want.
In terms of the free R set being 'contaminated', it shouldn't be a serious issue. That is to say, if you chose a free R set in P1 and chose a free R set in C2, then those two free R sets should be independent of each other. As long as you didn't try to pull one free R set into the other space group, you should be fine. [As a note, many people have discussed this previously and after many rounds of refinement, there shouldn't be any issue with a 'contaminated' free R set in any case, even if you didn't do it correctly].
I would like to add that the important point that I wanted to make is that if you have a pseudo two fold axis, those 'related' reflections are independent and should not be merged (even if they are similar in intensity). They will describe (potentially) real differences in those independent protomers in your asymmetric unit.
Best of luck, Tom
________________________________
From: Jorge Iulek
Sent: Friday, December 26, 2025 11:59 PM
To: Tom Peat ; [email protected]
Subject: Re: [phenixbb] transform C2 to P1 and the R-free set
Hi, Tom!
A bit more context.
I can process the data in C2 and, naturally, in P1.
In C2, indexes get a bit ugly. It goes, but some off the usual trend (smoothly lowering CC1/2 with resolution), it goes slight up and down. P1 "looks" "normal".
Yet, at refinement, at ~ the same advancement , C2 gives me
REMARK 3 R VALUE (WORKING SET) : 0.3719
REMARK 3 FREE R VALUE : 0.4331
and P1 gives me
REMARK 3 R VALUE (WORKING SET) : 0.2298
REMARK 3 FREE R VALUE : 0.2700
Naturally, C2 is double the volume, but its asymmetric unit has half (6,by the way) the number of molecules.
Brief density analyses (indexes), in any case (C2 and P1), show that most of the structure is already on due place. Possibly, further refinement will show up some subtle differences/difficulties (I think).
These indicators make me hypothesize that the 2 fold is not really crystallographic, but just nearly. I think that, though reflections intensities are not "completely" related , pairs of reflections related (their HKL indexes) by this 2 fold should be put in the same set. I recall this subject to have come up to forums before. By this way, one might ask how large might a deviation from a true 2-fold crystallographic be for the reflections not to share any further relation and so do not "contaminate" an Rfree set.
I am still looking around, possibly ccp4 pointless can help on this, but I must still study better.
Thank you,
Jorge
On 24/12/2025 19:39, Tom Peat wrote:
Hello Jorge,
If it is a pseudo two fold, then the reflections relating the pseudo two fold are not identical, so should not be scaled and merged together. Even if they are very close in value/ intensity, these intensities should not be merged (under almost all circumstances). The programs are pretty good about picking a fairly random set of free R reflections and setting the space group to P1 should allow everything to work as it should. There is nothing that should be done if it is a pseudo two fold.
If it is a crystallographic two fold, then you do need to process in the monoclinic space group to make sure everything works properly and an appropriate set of free R reflections are chosen.
I hope that helps, Tom
________________________________
From: Jorge Iulek mailto:[email protected]
Sent: Thursday, December 25, 2025 8:57 AM
To: Tom Peat mailto:[email protected]; [email protected]mailto:[email protected] mailto:[email protected]
Subject: Re: [phenixbb] transform C2 to P1 and the R-free set
Hi, Tom,
Thanks for your input.
Hello Jorge,
I'm not sure I understand your question, so this answer may not be appropriate, but: if you think the two fold is pseudo, why not just run with the P1 space group?
Setting the space group to P1 will allow the program you use to get an appropriate set of reflections for your free R set and your working set.
But I think in this case it would not know about the pseudo 2 so it will not care to include reflections related by this "nearly" 2 fold in the R-free set.
Well, one question here might be how much the 2 fold should depart from crystallographic to pseudo-crystallographic such that a (free) random free set would not be influenced by the pseudo 2 fold.
Cheers,
Jorge
Happy holidays, Tom
________________________________
From: Jorge Iulek mailto:[email protected]
Sent: Wednesday, December 24, 2025 10:58 PM
To: [email protected]mailto:[email protected] mailto:[email protected]
Subject: [phenixbb] transform C2 to P1 and the R-free set
Dear all,
I got a set of images (very recently) . Dials gives me:
+------------+--------------+--------+--------------+----------+-----------+-------------------------------------------+----------+-----------+
| Solution | Metric fit | rmsd | min/max cc | #spots |
lattice | unit_cell | volume |
cb_op |
|------------+--------------+--------+--------------+----------+-----------+-------------------------------------------+----------+-----------|
| * 2 | 0.3001 | 0.066 | 0.272/0.272 | 10685 | mC
| 212.43 81.92 133.04 90.00 113.12 90.00 | 2129290 | a-2*b,a,c |
| * 1 | 0 | 0.052 | -/- | 10559 | aP
| 81.95 113.64 133.10 68.50 89.98 69.13 | 1065523 | a,b,c |
+------------+--------------+--------+--------------+----------+-----------+-------------------------------------------+----------+-----------+
Many indications are that the 2-fold axis is not really
crystallographic, but a pseudo one. My point in this case is that
reflections related by this "almost" 2-fold axis should go into the same
set, free and work, this in the (new) P1 setting.
I could not devise clearly how to get a program to do this
automatically (say, flag reflections with this pseudo 2 related for
free -set ) . Is there? I understand cb_op is the real space operation,
possibly out of a crystallography package I can get that, but easier if
there is a crystallographic tool.
Any thoughts?
Thanks,
Jorge
(Season's Greetings)
_______________________________________________
phenixbb mailing list -- [email protected]mailto:[email protected]
To unsubscribe send an email to [email protected]mailto:[email protected]
Unsubscribe: phenixbb-leave@%(host_name)s
This email and any files transmitted with it may contain confidential information. If you believe you have received this email or any of its contents in error, please notify me immediately by return email and destroy this email. Do not use, disseminate, forward, print or copy any contents of an email received in error.
