- phenixbb - phenix-online.org

Re: [phenixbb] AutoMR and capsids
by SIPPEL,KATHERINE H 22 Oct '09

22 Oct '09

Dear Randy, > If you're sure that there are 3 full copies (and that you > couldn't, say, have six half-copies sitting on crystallographic > 2-folds), the best would be to search for full capsids. Now, for > each copy there are 60 different ways that the capsid could be > oriented, but the version of Phaser distributed with the latest > Phenix release should recognize the internal symmetry -- let us > know if it doesn't! I'm pretty sure its not actually full copies. The unit cell is 262.9, 262.9, 609.1 with an R32 space group. I'm downloading the latest release at the moment. > If it turns out that a capsid is sitting on a symmetry axis, > Phaser will probably reject it because of an outrageous number of > clashes. In that case, you should turn off the clash test, just > look for one copy, and then examine the overlapped solution to > see which part of the capsid is crystallographically unique. Thanks for the tips on the ASU, I wouldn't have thought of that. > Hi Katherine, > Back when I did this with HBV I had one capsid per > crystallographic AsU and I >used Amore (pre-phaser or phenix). I > searched with the icosohedral ASU >looking for 60 copies, and was > surprised how easily the solution fell out. If >you use this > strategy, you really just need one icos ASU per capsid and you > >can then expand the rest using standard icos symmetry. > I think the answer will be do what's easiest, and if it's 99% ID > it should >fall out pretty easily anyway. Bear in mind that > depending on what you've >done to these capsids the icos ASU's > may have slightly re-oriented (in my case >I had added a drug > that kept the icos ASU identical but had tilted the whole >thing > relative to the plane of the capsid). This may also guide your > strategy. > Good luck, > Christina Our standard protocol is Amore with CNS. When Amore failed I thought I might try something different. I have really enjoyed Phenix for my small proteins and thought that I might take advantage of that lovely NCS feature in phenix.refine. I just wasn't sure how to proceed within the AutoMR program. Obviously in the end I will go with what works. > PS- Isn't Mavis around in Florida somewhere, or are you even in > her lab? She >and Rob are a great resource.... I do work for Rob and Mavis and they are a fabulous resource. Best bosses ever (and no they won't see this, its just the truth). Thanks for the help, Katherine > On Oct 22 2009, SIPPEL,KATHERINE H wrote: > >> Dear all, >> >> I am currently trying to use AutoMR to do molecular replacement >> on an icosahedral virus capsid. It is a T=1 capsid so there are >> 60 copies of the monomer per capsid and based on the Matthews >> coefficient there are ~3 capsids in the unit cell. The search >> model is 99% identical (thank God) but I am unsure how to >> approach this. >> >> Do I used the 60-mer as the model and search for three copies? >> Do I use a monomer for the model and search for 180 copies? >> Should I break the model down to smaller pieces (i.e. a 30-mer >> and copies=6, or 12-mer and copies=15)? >> >> I would appreciate any help I can get on this one. >> >> Thanks, >> >> Katherine >> >> -- >> SIPPEL,KATHERINE H >> Ph. D. candidate >> Department of Biochemistry and Molecular Biology >> College of Medicine >> University of Florida >> >> _______________________________________________ >> phenixbb mailing list >> phenixbb(a)phenix-online.org >> http://www.phenix-online.org/mailman/listinfo/phenixbb >> > > _______________________________________________ > phenixbb mailing list > phenixbb(a)phenix-online.org > http://www.phenix-online.org/mailman/listinfo/phenixbb > > -- SIPPEL,KATHERINE H Ph. D. candidate McKenna Lab Department of Biochemistry and Molecular Biology College of Medicine University of Florida

1 0

Another questions: water for adp and partial occupancy
by Young-Jin Cho 22 Oct '09

22 Oct '09

Hi, I thought we should put water molecules as isotropic and full occupancies. However, I saw different cases. So I am wondering the reasons for putting waters as anisotropic and with partial occupancies in an active sites. Does any one have this kind of experience or can you explain the advantage or disadvante of this types of strategy? Thanks in advance, Young-Jin Cho

2 1

AutoMR and capsids
by SIPPEL,KATHERINE H 22 Oct '09

22 Oct '09

Dear all, I am currently trying to use AutoMR to do molecular replacement on an icosahedral virus capsid. It is a T=1 capsid so there are 60 copies of the monomer per capsid and based on the Matthews coefficient there are ~3 capsids in the unit cell. The search model is 99% identical (thank God) but I am unsure how to approach this. Do I used the 60-mer as the model and search for three copies? Do I use a monomer for the model and search for 180 copies? Should I break the model down to smaller pieces (i.e. a 30-mer and copies=6, or 12-mer and copies=15)? I would appreciate any help I can get on this one. Thanks, Katherine -- SIPPEL,KATHERINE H Ph. D. candidate Department of Biochemistry and Molecular Biology College of Medicine University of Florida

3 2

Partially occupied SO4 and waters
by Chris Waddling 20 Oct '09

20 Oct '09

I have a partially occupied sulfate ion that is refining to 55% along with one of two arginine side chains. Associated with the second conformation of the arginine, there is positive density for three partially occupied waters. In Shelxl, I can use the PART identifier to specify that the sulfate ion and the waters are not present simultaneously and are associated with their respective Arg side chain conformation. Is there a way to do this in Phenix? Thanks, Chris -- Dr. Christopher A. Waddling, Ph.D. University of California at San Francisco

3 2

Re: [phenixbb] Molprobity percentile statistic
by hari jayaram 20 Oct '09

20 Oct '09

Thanks Nat and Pavel for your emails, Thanks for the suggestion of using phenix.polygon. The comparative analysis of my structure relative to others at my resolution was what I was after. And thats the reason I liked the molprobity statistics expressed as percentile since it helped me realize how far I have to go relative to other structures at the PDB. If its not too difficult can I cast my vote to report it inside the phenix GUI. Its just that running the full analysis inside of phenix is significantly faster than running it on the duke webservers. In molprobity If I remember correctly the overall score is some weighted average, and the molprobity paper also cautions against taking it too literally. Also If I remember the molprobity percentiles are absolute and not only composed of structures at similar resolutions . Regardless thanks for the hint about phenix.polygon . I just looked at my phenix.polygon report and I am afraid its a little asymmetric. Hari On Mon, Oct 19, 2009 at 9:58 PM, hari jayaram <harijay(a)gmail.com> wrote: > Thanks Nat and Pavel for your emails, > Thanks for the suggestion of using phenix.polygon. The comparative > analysis of my structure relative to others at my resolution was what > I was after. > > And thats the reason I liked the molprobity statistics expressed as > percentile since it helped me realize how far I have to go relative to > other structures at the PDB. If its not too difficult can I cast my > vote to report it inside the phenix GUI. Its just that running the > full analysis inside of phenix is significantly faster than running it > on the duke webservers. > > In molprobity If I remember correctly the overall score is some > weighted average, and the molprobity paper also cautions against > taking it too literally. Also If I remember the molprobity percentiles > are absolute and not only composed of structures at similar > resolutions . > > Regardless thanks for the hint about phenix.polygon . > I am attaching a snapshot of my polygon ..I an afraid its a little > asymmetric. > > Hari > > On Mon, Oct 19, 2009 at 7:48 PM, Pavel Afonine <PAfonine(a)lbl.gov> wrote: >> Hi Hari, >> >> if you run phenix.model_vs_data you will get those numbers (namely: >> Ramachandran outliers, clashscores, CB-deviations; although you will not >> get "Molprobity score" simply because I don't know what it is and so I >> didn't add it): >> >> phenix.model_vs_data model.pdb data.mtz >> >> Nat: every time time you run model_vs_data for whatever purpose (getting >> fmodel, etc...) you get an object that contains all these numbers. >> >> Pavel. >> >> >> On 10/19/09 3:32 PM, hari jayaram wrote: >>> Hi i just started using the phenix GUI front-end to Molprobity . Its >>> really nice the way it plugs into coot and displays all the statistics >>> etc. >>> Is there any place that it also reports the percentile clashscore and >>> percentile Molprobity score for overall model quality a la the >>> webserver at molprobity.biochem.duke.edu >>> >>> Thanks >>> Hari >>> _______________________________________________ >>> phenixbb mailing list >>> phenixbb(a)phenix-online.org >>> http://www.phenix-online.org/mailman/listinfo/phenixbb >>> >> _______________________________________________ >> phenixbb mailing list >> phenixbb(a)phenix-online.org >> http://www.phenix-online.org/mailman/listinfo/phenixbb >> >

2 1

Molprobity percentile statistic
by hari jayaram 19 Oct '09

19 Oct '09

Hi i just started using the phenix GUI front-end to Molprobity . Its really nice the way it plugs into coot and displays all the statistics etc. Is there any place that it also reports the percentile clashscore and percentile Molprobity score for overall model quality a la the webserver at molprobity.biochem.duke.edu Thanks Hari

3 2

Re: [phenixbb] pseudo-merohedral twinning or pseudo-symmetry?
by Gino Cingolani 16 Oct '09

16 Oct '09

Laurie, I had a similar problem. Scaling statistics in P212121 were OK, and - of course - no twinning was detected ... but nothing made sense during refinement. Looking at the data more carefully I found that if one rescales in P21, the b-angle was stably refined to 90.20 with ~30% better Rsym than in orthorhombic. When running xtriage on the P21 data, I found the data were indeed "pseudomerohedrally-twinned" with twinning fraction ~0.38. Good luck! Gino

3 2

pseudo-merohedral twinning or pseudo-symmetry?
by Laurie Betts 16 Oct '09

16 Oct '09

Hi - We have a data set that seems to process well in primitive orthorhombic. a=80, b=87, c=114. The data has overall Rsymm around 10% to 2.6 Angstrom, where I/sigI > 2, and Chi-squared is around 1.4. Phenix.xtriage shows some pathology in intensity statistics such that the L-test indicates <|L|> = 0.414 and <L^2> = 0.241, with the L-test Z score = 7.334. The Wilson ratio for acentric refls is 1.774. According the xtriage, There were no "pseudomerohedral" twin laws found, and the comments at the end say "As there are no twin laws possible given the crystal symmetry, there could be a number of reasons......" Is it more likely that this is a case of pseudo-symmetry, or would one not expect the intensity statistics to have the values they do in that case. Also, is the absence of twin laws defined by the fact that proper merohedral twinning is not possible for P222 space groups? Still confused, Laurie Betts

2 1

autobuild
by Colin Levy 16 Oct '09

16 Oct '09

Hi, can autobuild be setup to use kicked maps when starting from a molecular replacement model? Thanks Colin

2 1

Re: [phenixbb] grouped occupancy definitions
by Dr. Mark Mayer 15 Oct '09

15 Oct '09

Hi Pavel, I'm digging in docs and list serv but could not find out, and figured you probably know faster. I need to make an occ.params file for some groups of non adjacent side chains which interact with each other, and associated waters. 1) How do I select ACs 2) How do I constrain water occupancies so that their Q + that of coupled protein atomes does not exceed 1. If I match the AC flags for the waters and side chains will this take care of it. e.g. if residues 2 and 21 have coupled ACs and waters 101 and 102 move in when residue 21 is in one of its ACs how would I define below? refinement { refine { occupancies { constrained_group { selection = chain A and resseq 2 selection = chain A and resseq 21 selection = chain W and resseq 101 selection = chain W and resseq 102 } } Thanks !!! -- +++++++++++++++++++++++++++++++++++++++++++++++++++++++ Mark Mayer Ph.D.

2 2