Re: [ccp4bb] on NCS restraint

[Victor Lamzin]

Dear Reza, > have the results been published? Not all of them. The first implementation of NCS-restraints in ARP/wARP's model building at a resolution of 2.3 A and lower was published: Wiegels T, Lamzin VS. Use of noncrystallographic symmetry for automated model building at medium to low resolution. (2012) Acta Crystallogr D Biol Crystallogr. 68, 446-453. PMID: 22505265 Subsequent advances and applicability to data up to 1.5 A is still to be published; however it is now the default setting of ARP/wARP. Victor On 21/04/2015 15:32, Reza Khayat wrote: Hi,   I have to agree with Steven. I was too haste in my reply to the initial e-mail. As Steven put it, it is best to run multiple experiments at the same time and identify which produces the best result.   On a follow up to Victor Lamin’s reply, have the results been published?   Best wishes, Reza     Reza Khayat, PhD Assistant Professor Department of Chemistry City College of New York New York, NY 10031 http://www.khayatlab.org/ 212-650-6070   From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Victor Lamzin Sent: Tuesday, April 21, 2015 9:07 AM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] on NCS restraint   Hi all, We have carried out a large-scale test of the use of Refmac's NCS-restraints during model building with ARP/wARP. We have found advantageous to have such restraints turned on with data resolution extending to as high as 1.5 A. Victor On 21/04/2015 14:46, Sheriff, Steven wrote: All:   I strongly disagree with Reza’s suggestion that one should abandon NCS at better than 2.5 Å resolution (or even Herman’s suggestion at better than 2.0 Å resolution). Either of these may be true in any particular case, BUT one should do the experiment – Run parallel refinements from the same starting model with and without NCS restraints and compare R-free, the gap between R-free and R-work, and particular places where one knows or suspects that the local geometry is different, before deciding to abandon NCS restraints. This was certainly true in the bad old days when “loose” (as opposed to strict) NCS restraints were used and “bound” each chain more-or-less to a single chain’s geometry, even though one was refining all extant chains. Using so-called “loose” NCS restraints, I once had a loop pulled out of electron density during refinement and the tip moved ~6 Å!   However, for the last 5 years or so, BUSTER, which I use, and REFMAC (and presumably PHENIX) have used LSSR (local secondary structure restraints) where the maximum “pull” to uniformity tops out at a certain value. I have not rigorously followed my own advice above to run parallel refinements, but I have yet to find a case where LSSR-type NCS restraints have “hurt” the refinement down to at least ~1.5 Å resolution.   To give credit where it is due, Oliver Smart, who implemented LSSR in BUSTER, attributed the concept to George Sheldrick.   Steven   ------------------------------ Date:    Mon, 20 Apr 2015 10:38:27 +0000 From:    Reza Khayat <rkha...@ccny.cuny.edu> Subject: Re: [ccp4bb] on NCS restraint   Hi,   The purpose of NCS is to reduce the degrees of freedom in order to avoid over refinement -not only to expedite refinement. Strict or restrained NCS should be applied at lower resolutions (<2.7Å) or data completeness. Forgo NCS If you have a complete and better than 2.5Å dataset. Also, you can define the regions where NCS is applied and thus avoid loops/regions where the NCS is violated.   Best wishes, Reza   Reza Khayat, PhD Assistant Professor City College of New York 160 Convent Ave, MR-1135 New York, NY 10031 (212) 650-6070 rkha...@ccny.cuny.edu<mailto:rkha...@ccny.cuny.edu>   ------------------------------ On Apr 20, 2015, at 4:01 AM, herman.schreu...@sanofi.com<mailto:herman.schreu...@sanofi.com> wrote:   Dear Smith,   There used to be something called “strict NCS“ which meant that instead of many identical subunits, only one “average” subunit was refined, which would speed up the refinement significantly, at the expense of requiring that all subunits are exactly identical.   I do not think that this option is used anymore and most refinement programs would require NCS related subunits to be similar, but not identical to each other. As Robbie Joosten pointed at, this can help a lot, especially when you do not have high resolution data. So for data with better than 2.0 Å resolution, including NCS restraints would probably not make a big difference, but otherwise I would switch them on.   Best, Herman   ------------------------------ Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Smith Liu Gesendet: Freitag, 17. April 2015 06:02 An: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> Betreff: Re: [ccp4bb] on NCS restraint   Dear Jurgen,   My understanding is that NCS restraint can significantly enhance the speed of calculation, but considering the subunits even with the eactly same sequence may not be identical, to have NCS restraint may be not necessary or may be not good for the refinement, am I right?   Smith   At 2015-04-17 09:09:05, "Jurgen Bosch" <jbos...@jhu.edu<mailto:jbos...@jhu.edu>> wrote:   yes. Have two sets of NCS operators one that describe the four subunits and one describing the two subunits. If during the refinement of your structure you should find out that the subunits are not identical to each other you can relax the NCS weights.   Jürgen ...................... Jürgen Bosch Johns Hopkins University Bloomberg School of Public Health Department of Biochemistry & Molecular Biology Johns Hopkins Malaria Research Institute 615 North Wolfe Street, W8708 Baltimore, MD 21205 Office: +1-410-614-4742<tel:%2B1-410-614-4742> Lab:      +1-410-614-4894<tel:%2B1-410-614-4894> Fax:      +1-410-955-2926<tel:%2B1-410-955-2926> http://lupo.jhsph.edu<http://lupo.jhsph.edu/>   ------------------------------ On Apr 16, 2015, at 9:02 PM, Smith Lee <00000459ef8548d5-dmarc-requ...@jiscmail.ac.uk<mailto:00000459ef8548d5-dmarc-requ...@jiscmail.ac.uk>> wrote:     Dear All,   If a protein contains 6 subunits, 4 subunits from the same sequence (subunit A, B, C, D all from the same sequence), each of the 2 other subunits from 2 diffrent sequences (subunit E from the second sequence, subunit F from the third sequence), in this situation should I use NCS restraint or not?   If my protein contains 2 subunits, both of the 2 subunits composed of the eaxctly same sequence, however supposing the 2 subunits have a little diffrent conformation, in this situation should we use NCS retraint or not?   Smith   ------------------------------   Date:    Mon, 20 Apr 2015 14:18:19 +0200 From:    Robbie Joosten <robbie_joos...@hotmail.com> Subject: Re: AW: [ccp4bb] on NCS restraint   Hi Herman,   Strictncs is still used for viral capsids and other high NCS structures. It works very well in Refmac as long as your MTRIX records (in the PDB file) are correct and you have the identity MTRIX as well. The keyword is simply 'strictncs'. You can even combine strict and local NCS if you want in Refmac (not sure you should want to do that, but you can).   Shameless plug: PDB_REDO automatically deals with strict NCS if you work on your own structures. PDB entries with strict NCS are still checked manually because there used to be a lot of annotation errors with respect to strict NCS. It is much less common now, but I still find the odd new one :(   Cheers, Robbie   ------------------------------   Date: Mon, 20 Apr 2015 08:01:26 +0000 From: herman.schreu...@sanofi.com Subject: [ccp4bb] AW: [ccp4bb] on NCS restraint To: CCP4BB@JISCMAIL.AC.UK   Dear Smith, There used to be something called “strict NCS“ which meant that instead of many identical subunits, only one “average” subunit was refined, which  would speed up the refinement significantly, at the expense of requiring that all subunits are exactly identical.   I do not think that this option is used anymore and most refinement programs would require NCS related subunits to be similar, but not identical  to each other. As Robbie Joosten pointed at, this can help a lot, especially when you do not have high resolution data. So for data with better than 2.0 Å resolution, including NCS restraints would probably not make a big difference, but otherwise I would  switch them on.   Best, Herman ------------------------------ Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Smith Liu Gesendet: Freitag, 17. April 2015 06:02 An: CCP4BB@JISCMAIL.AC.UK Betreff: Re: [ccp4bb] on NCS restraint   Dear Jurgen,   My understanding is that NCS restraint can significantly enhance the speed of calculation, but considering the subunits even with the  eactly same sequence may not be identical, to have NCS restraint may be not necessary or may be not good for the refinement, am I right?   Smith   At 2015-04-17 09:09:05, "Jurgen Bosch" <jbos...@jhu.edu> wrote:   yes.   Have two sets of NCS operators one that describe the four subunits and one describing the two subunits. If during the refinement of your  structure you should find out that the subunits are not identical to each other you can relax the NCS weights.   Jürgen   ......................   Jürgen Bosch Johns Hopkins University Bloomberg School of Public Health Department of Biochemistry & Molecular Biology Johns Hopkins Malaria Research Institute 615 North Wolfe Street, W8708 Baltimore, MD 21205 Office: +1-410-614-4742 Lab:      +1-410-614-4894 Fax:      +1-410-955-2926 http://lupo.jhsph.edu     ------------------------------ On Apr 16, 2015, at 9:02 PM, Smith Lee <00000459ef8548d5-dmarc-requ...@jiscmail.ac.uk> wrote:   Dear All,   If a protein contains 6 subunits, 4 subunits from the same sequence (subunit A, B, C, D all from the same sequence), each of the 2 other  subunits from 2 diffrent sequences (subunit E from the second sequence, subunit F from the third sequence), in this situation should I use NCS restraint or not?   If my protein contains 2 subunits, both of the 2 subunits composed of the eaxctly same sequence, however supposing the 2 subunits have  a little diffrent conformation, in this situation should we use NCS retraint or not?   Smith                    This message (including any attachments) may contain confidential, proprietary, privileged and/or private information. The information is intended to be for the use of the individual or entity designated above. If you are not the intended recipient of this message, please notify the sender immediately, and delete the message and any attachments. Any disclosure, reproduction, distribution or other use of this message or any attachments by an individual or entity other than the intended recipient is prohibited.