Re: [ccp4bb] large R-Rfree difference in "final" structure
Dear Careina, As other have mentioned, my feeling is that something is suspicious. However the problem might be very difficult or impossible to catch and other structures have been deposited with similar R/Rfrees... What I would check in addition to what has been mentioned, is if some twinning is present. With say a twin-factor of 10-20% processing may proceed sort of ok and refinement as well with a somewhat higher Rfree. You may also suffer from disorder. You mention that you model is complete: does that mean that the complete sequence of your protein has been fitted, or does this mean that no more positive difference density is left? If a significant portion of your protein is not visible, this may still compromise your Rfactors. In the latter case, I don't think you can do much about it, except offering this as an explanation for your high R/Rfree. Best, Herman From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Careina Edgooms Sent: Wednesday, July 13, 2011 5:39 PM To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] large R-Rfree difference in "final" structure Dear ccp4 bulletin board I just have a slight concern regarding my Rwork Rfree difference. I have a structure that I have solved. I am reasonably content that it is complete because it has refined well, it no longer has bad geometries and contacts and all the rotamers, ramachandra, bond lengths etc are good. It gives favourable scores on molprobity and procheck. My only concern is the R factor difference. The resolution of the structure is 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather high. Should I be concerned? During refinement Rfree only drops from about 0.36 to 0.33 while the R factor drops from 0.31 to 0.24.. I have removed automatic weighting in refmac in order to constrain my bond lengths and angles during a couple of rounds of refinement. This did not have any effect on the R factors, however. I am fairly content that the space group I have chosen is correct so I am not sure what else could cause the big difference in R factors? There is no twinning. Can I be satisfied that my structure is correct despite the high R free or should I be doing other checks/ trying other things before I can submit this structure? Thank you for any help Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
Hi, Ian Tickle has asked the most pertinent question - there are many reasons for highish FreeR factors. Only some are due to errors or missing molecules or other problems. How many "Free" reflections are there? When the sample is very small (< 500 say) you can get different results from a different Free R set. And remember that the numbers you quote are overall values, It is worth looking at the plot of R and Free R against resolution. They should stay reasonably parallel, but if the data becomes very weak at the resolution limit the the FreeR can shoot up relative to the overall Rvalue. I use COOT validation to search the difference maps for peaks and holes. If there is something missing from your model you should see it in that list.. Eleanor On Wed, 13 Jul 2011 20:37:51 +0100, Matthias Zebisch wrote: > Hi Careina, > > our lab we once had the problem, that the asymmetric unit contained 8 > molecules, > whereas 7 had only been modeled. Somehow the 8th monomer had evaded > detection. > So be careful not to miss density. > > Matthias > > > On 7/13/2011 7:54 PM, Robbie Joosten wrote: >> Hi Careina, >> >> >> >> Assuming you don't suffer from a very poor data parameter ratio that >> would lead to such a large R-free/R, you need to improve your refinement. >> If you have NCS you should use local NCS restraints. You could also try >> jelly-body restraints, although they may not work at your resolution. >> >> >> >> Cheers, >> >> Robbie >> >> ____________ >>> Date: Wed, 13 Jul 2011 08:38:38 -0700 >>> From: careinaedgo...@yahoo.com >>> Subject: [ccp4bb] large R-Rfree difference in "final" structure >>> To: CCP4BB@JISCMAIL.AC.UK >>> >>> Dear ccp4 bulletin board >>> >>> I just have a slight concern regarding my Rwork Rfree difference. I >>> have a structure that I have solved. I am reasonably content that it is >>> complete because it has refined well, it no longer has bad geometries >>> and contacts and all the rotamers, ramachandra, bond lengths etc are >>> good. It gives favourable scores on molprobity and procheck. My only >>> concern is the R factor difference. The resolution of the structure is >>> 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which >>> seems to me to be rather high. Should I be concerned? >>> >>> During refinement Rfree only drops from about 0.36 to 0.33 while the R >>> factor drops from 0.31 to 0.24.. I have removed automatic weighting in >>> refmac in order to constrain my bond lengths and angles during a couple >>> of rounds of refinement. This did not have any effect on the R factors, >>> however. I am fairly content that the space group I have chosen is >>> correct so I am not sure what else could cause the big difference in R >>> factors? There is no twinning. >>> >>> Can I be satisfied that my structure is correct despite the high R free >>> or should I be doing other checks/ trying other things before I can >>> submit this structure? >>> >>> Thank you for any help >>> Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
Hi Careina, in our lab we once had the problem, that the asymmetric unit contained 8 molecules, whereas 7 had only been modeled. Somehow the 8th monomer had evaded detection. So be careful not to miss density. Matthias On 7/13/2011 7:54 PM, Robbie Joosten wrote: Hi Careina, Assuming you don't suffer from a very poor data parameter ratio that would lead to such a large R-free/R, you need to improve your refinement. If you have NCS you should use local NCS restraints. You could also try jelly-body restraints, although they may not work at your resolution. Cheers, Robbie Date: Wed, 13 Jul 2011 08:38:38 -0700 From: careinaedgo...@yahoo.com Subject: [ccp4bb] large R-Rfree difference in "final" structure To: CCP4BB@JISCMAIL.AC.UK Dear ccp4 bulletin board I just have a slight concern regarding my Rwork Rfree difference. I have a structure that I have solved. I am reasonably content that it is complete because it has refined well, it no longer has bad geometries and contacts and all the rotamers, ramachandra, bond lengths etc are good. It gives favourable scores on molprobity and procheck. My only concern is the R factor difference. The resolution of the structure is 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather high. Should I be concerned? During refinement Rfree only drops from about 0.36 to 0.33 while the R factor drops from 0.31 to 0.24.. I have removed automatic weighting in refmac in order to constrain my bond lengths and angles during a couple of rounds of refinement. This did not have any effect on the R factors, however. I am fairly content that the space group I have chosen is correct so I am not sure what else could cause the big difference in R factors? There is no twinning. Can I be satisfied that my structure is correct despite the high R free or should I be doing other checks/ trying other things before I can submit this structure? Thank you for any help Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
In my experience, the most common cause of this is the "locally correct, globally wrong" issue. You could be in the wrong space group (i.e P21212 when you should be in P212121). Another possibility is missing some element of local symmetry. Is the packing in the unit cell appropriate or are you underpacked? Do you have pseudosymmetry, which can mask the effects of merohedral twinning on certain measures? Adrian. On 14 Jul 2011, at 04:54, Robbie Joosten wrote: > Hi Careina, > > > > Assuming you don't suffer from a very poor data parameter ratio that would > lead to such a large R-free/R, you need to improve your refinement. If you > have NCS you should use local NCS restraints. You could also try jelly-body > restraints, although they may not work at your resolution. > > > > Cheers, > > Robbie > > >> Date: Wed, 13 Jul 2011 08:38:38 -0700 >> From: careinaedgo...@yahoo.com >> Subject: [ccp4bb] large R-Rfree difference in "final" structure >> To: CCP4BB@JISCMAIL.AC.UK >> >> Dear ccp4 bulletin board >> >> I just have a slight concern regarding my Rwork Rfree difference. I >> have a structure that I have solved. I am reasonably content that it is >> complete because it has refined well, it no longer has bad geometries >> and contacts and all the rotamers, ramachandra, bond lengths etc are >> good. It gives favourable scores on molprobity and procheck. My only >> concern is the R factor difference. The resolution of the structure is >> 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which >> seems to me to be rather high. Should I be concerned? >> >> During refinement Rfree only drops from about 0.36 to 0.33 while the R >> factor drops from 0.31 to 0.24.. I have removed automatic weighting in >> refmac in order to constrain my bond lengths and angles during a couple >> of rounds of refinement. This did not have any effect on the R factors, >> however. I am fairly content that the space group I have chosen is >> correct so I am not sure what else could cause the big difference in R >> factors? There is no twinning. >> >> Can I be satisfied that my structure is correct despite the high R free >> or should I be doing other checks/ trying other things before I can >> submit this structure? >> >> Thank you for any help >> Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
Hi Careina, Assuming you don't suffer from a very poor data parameter ratio that would lead to such a large R-free/R, you need to improve your refinement. If you have NCS you should use local NCS restraints. You could also try jelly-body restraints, although they may not work at your resolution. Cheers, Robbie > Date: Wed, 13 Jul 2011 08:38:38 -0700 > From: careinaedgo...@yahoo.com > Subject: [ccp4bb] large R-Rfree difference in "final" structure > To: CCP4BB@JISCMAIL.AC.UK > > Dear ccp4 bulletin board > > I just have a slight concern regarding my Rwork Rfree difference. I > have a structure that I have solved. I am reasonably content that it is > complete because it has refined well, it no longer has bad geometries > and contacts and all the rotamers, ramachandra, bond lengths etc are > good. It gives favourable scores on molprobity and procheck. My only > concern is the R factor difference. The resolution of the structure is > 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which > seems to me to be rather high. Should I be concerned? > > During refinement Rfree only drops from about 0.36 to 0.33 while the R > factor drops from 0.31 to 0.24.. I have removed automatic weighting in > refmac in order to constrain my bond lengths and angles during a couple > of rounds of refinement. This did not have any effect on the R factors, > however. I am fairly content that the space group I have chosen is > correct so I am not sure what else could cause the big difference in R > factors? There is no twinning. > > Can I be satisfied that my structure is correct despite the high R free > or should I be doing other checks/ trying other things before I can > submit this structure? > > Thank you for any help > Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
Careina, I recommend to compare the quality of your data (Rmerge) to that of an average data set of same resolution. Do you have meaningful data to 2.3A resolution? Another possibility is the anisotropicity of your data. Try this server http://services.mbi.ucla.edu/anisoscale/, if the anisotropicity is significant. Alex On Jul 13, 2011, at 8:38 AM, Careina Edgooms wrote: > Dear ccp4 bulletin board > > I just have a slight concern regarding my Rwork Rfree difference. I have a > structure that I have solved. I am reasonably content that it is complete > because it has refined well, it no longer has bad geometries and contacts and > all the rotamers, ramachandra, bond lengths etc are good. It gives favourable > scores on molprobity and procheck. My only concern is the R factor > difference. The resolution of the structure is 2.3A. The R factor is 0.24 > after refinement but the Rfree is 0.33 which seems to me to be rather high. > Should I be concerned? > > During refinement Rfree only drops from about 0.36 to 0.33 while the R factor > drops from 0.31 to 0.24.. I have removed automatic weighting in refmac in > order to constrain my bond lengths and angles during a couple of rounds of > refinement. This did not have any effect on the R factors, however. I am > fairly content that the space group I have chosen is correct so I am not sure > what else could cause the big difference in R factors? There is no twinning. > > Can I be satisfied that my structure is correct despite the high R free or > should I be doing other checks/ trying other things before I can submit this > structure? > > Thank you for any help > Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
Hi Careina, Your starting R/Rfree values seem just fine. The reason for the large disparity after refinement is because the weighting factor is way too large during the first rounds of refinement. For each round of refinement, have a few different runs which test weighting factors between 0.01-0.2. I have found that during the first round of refinement, running sequential rounds of 1 cycle refinements with gradually increasing weighting terms (0.01, 0.015, 0.02, 0.025, etc), works really well. This is a similar approach to how BALBES does its initial refinement. The smaller the weighting factor, the smaller the amount of change you will see between R/Rfree... at the expense of decreasing the amount that R/Rfree go down during each refinement. Another thing you could try if you are using Refmac is to switch from "Simple" to "Babinet" scaling. Hope this helps --- Greg Costakes PhD Candidate Department of Structural Biology Purdue University Hockmeyer Hall, Room 320 240 S. Martin Jischke Drive, West Lafayette, IN 47907 - Original Message - From: "Careina Edgooms" To: CCP4BB@JISCMAIL.AC.UK Sent: Wednesday, July 13, 2011 11:38:38 AM Subject: [ccp4bb] large R-Rfree difference in "final" structure Dear ccp4 bulletin board I just have a slight concern regarding my Rwork Rfree difference. I have a structure that I have solved. I am reasonably content that it is complete because it has refined well, it no longer has bad geometries and contacts and all the rotamers, ramachandra, bond lengths etc are good. It gives favourable scores on molprobity and procheck. My only concern is the R factor difference. The resolution of the structure is 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather high. Should I be concerned? During refinement Rfree only drops from about 0.36 to 0.33 while the R factor drops from 0.31 to 0.24.. I have removed automatic weighting in refmac in order to constrain my bond lengths and angles during a couple of rounds of refinement. This did not have any effect on the R factors, however. I am fairly content that the space group I have chosen is correct so I am not sure what else could cause the big difference in R factors? There is no twinning. Can I be satisfied that my structure is correct despite the high R free or should I be doing other checks/ trying other things before I can submit this structure? Thank you for any help Careina
Re: [ccp4bb] large R-Rfree difference in "final" structure
What leads you to believe Rfree is too high? - without knowing your data/parameter ratio it's impossible to tell. If you look at our paper Acta Cryst. (1998). D54, 547-557 (see Fig 1), then calculate the ratio y = Rfree/Rwork: in this case it is 0.33/0.24 = 1.375. Now look at the curve labelled 'a=2' (for normal restrained isotropic refinement) for the theoretically expected ratio and the green dots for the observed ratios (structures in the PDB at resolutions between 2.0 and 2.5 A). Then you need to calculate the ratio x = (no of atoms in the a.u.) / (no of reflections in the working set) and see where the point (x,y) falls relative to the curve and the green dots. If it's on or near the curve and in the middle of the green dots it's fine, if not then we need to work out why. -- Ian On Wed, Jul 13, 2011 at 4:38 PM, Careina Edgooms wrote: > Dear ccp4 bulletin board > > I just have a slight concern regarding my Rwork Rfree difference. I have a > structure that I have solved. I am reasonably content that it is complete > because it has refined well, it no longer has bad geometries and contacts > and all the rotamers, ramachandra, bond lengths etc are good. It gives > favourable scores on molprobity and procheck. My only concern is the R > factor difference. The resolution of the structure is 2.3A. The R factor is > 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather > high. Should I be concerned? > > During refinement Rfree only drops from about 0.36 to 0.33 while the R > factor drops from 0.31 to 0.24.. I have removed automatic weighting in > refmac in order to constrain my bond lengths and angles during a couple of > rounds of refinement. This did not have any effect on the R factors, > however. I am fairly content that the space group I have chosen is correct > so I am not sure what else could cause the big difference in R factors? > There is no twinning. > > Can I be satisfied that my structure is correct despite the high R free or > should I be doing other checks/ trying other things before I can submit this > structure? > > Thank you for any help > Careina >
Re: [ccp4bb] large R-Rfree difference in "final" structure
Hi Careina, I just have a slight concern regarding my Rwork Rfree difference. I have a > structure that I have solved. I am reasonably content that it is complete > because it has refined well, it no longer has bad geometries and contacts > and all the rotamers, ramachandra, bond lengths etc are good. It gives > favourable scores on molprobity and procheck. My only concern is the R > factor difference. The resolution of the structure is 2.3A. The R factor is > 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather > high. Should I be concerned? > looking at all similar to yours structures in PDB (*): phenix.r_factor_statistics 2.3 Histogram of Rwork for models in PDB at resolution 2.20-2.40 A: 0.126 - 0.147 : 29 0.147 - 0.168 : 253 0.168 - 0.189 : 958 0.189 - 0.210 : 1649 0.210 - 0.232 : 1440 0.232 - 0.253 : 549 <<< your structure 0.253 - 0.274 : 125 0.274 - 0.295 : 12 0.295 - 0.316 : 1 0.316 - 0.337 : 4 Histogram of Rfree for models in PDB at resolution 2.20-2.40 A: 0.165 - 0.188 : 31 0.188 - 0.211 : 223 0.211 - 0.235 : 860 0.235 - 0.258 : 1733 0.258 - 0.281 : 1508 0.281 - 0.304 : 542 0.304 - 0.328 : 97 0.328 - 0.351 : 19 <<< your structure 0.351 - 0.374 : 6 0.374 - 0.397 : 1 Histogram of Rfree-Rwork for all model in PDB at resolution 2.20-2.40 A: 0.001 - 0.011 : 53 0.011 - 0.021 : 172 0.021 - 0.031 : 524 0.031 - 0.041 : 960 0.041 - 0.050 : 1264 0.050 - 0.060 : 1050 0.060 - 0.070 : 592 0.070 - 0.080 : 224 0.080 - 0.090 : 114 0.090 - 0.100 : 67 <<< your structure Number of structures considered: 5020 (*): *Crystallographic model quality at a glance*. Acta Cryst. D65, 297-300 It seems like you should be concerned about the high Rfree and Rfree-Rwork. The second question about how to address it is program specific in some sense and I hope Refmac experts will address it. Pavel.
[ccp4bb] large R-Rfree difference in "final" structure
Dear ccp4 bulletin board I just have a slight concern regarding my Rwork Rfree difference. I have a structure that I have solved. I am reasonably content that it is complete because it has refined well, it no longer has bad geometries and contacts and all the rotamers, ramachandra, bond lengths etc are good. It gives favourable scores on molprobity and procheck. My only concern is the R factor difference. The resolution of the structure is 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather high. Should I be concerned? During refinement Rfree only drops from about 0.36 to 0.33 while the R factor drops from 0.31 to 0.24.. I have removed automatic weighting in refmac in order to constrain my bond lengths and angles during a couple of rounds of refinement. This did not have any effect on the R factors, however. I am fairly content that the space group I have chosen is correct so I am not sure what else could cause the big difference in R factors? There is no twinning. Can I be satisfied that my structure is correct despite the high R free or should I be doing other checks/ trying other things before I can submit this structure? Thank you for any help Careina
