[ccp4bb] AW: [ccp4bb] Problem with Rfactor
Dear Sudarshan, There are quite a few point to consider in MR: -How good is your model? At >50% sequence identity with your protein, it is probably ok, around 25% sequence identity it may or may not work. If your protein is of the same protein family with a similar biological function it is probably ok. -How many protein monomers should I expect? Proteins have in general between 25 and 75% solvent, so e.g. with the Matthews program one can estimate how many molecules might be in the asymmetric unit. Conversely, if your search model contains more than one molecule in the asymmetric unit, you have to delete the additional copies, otherwise many MR programs will fail. -If your search model has low homology, you may want to trim away surface loops which are missing or likely different in your protein. You may also want to change side chains that are different to Ala, so they won’t cause problems. -Space group: As Eleanor pointed out, there are many hexagonal space groups and you have to try them all. -Important: before you continue, check that both the mtz file and the pdb file you continue with have the correct space group, found by the MR program. Other wise, you have to reset the space group, e.g. with cad or pdbset. -A high Rfactor just after MR may not be a problem, probably quite a few residues have to be mutated from the search sequence to your sequence and many loops may need rebuilding too. Important is that you recognize some of your residues in the electron density maps. -The thing to do is to rebuild your model and run a few rounds of refinement. If the Rfactors do not go down, you have a problem, but otherwise you can just continue. Good luck! Herman Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Sudarshan Murthy Gesendet: Dienstag, 29. April 2014 06:11 An: CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] Problem with Rfactor Dear All, I am very new to the field of crystallography, I have a few questions which are very basic and getting input from this forum would help me a lot. Firstly I am trying to solve a structure using MR Molrep. In the result the Rfactors are really high how do I reduce the same?? ..When I tried with Phaser I didnt get any solution at all.. Will der be a problem with the space group?? like the model is in monoclinic and the data which I have is processed in hexagonal. Secondly while using hexagonal space group should I search for only one monomer in Molrep?? These are very basic questions if anybody could help me out with this , I would be very happy for the same. Sudarshan .N. Murthy Crystallography Division Bangalore
Re: [ccp4bb] Problem with Rfactor
We need more information. First if the pointgroup is hexagonal are you searching al likely SPACEgroup - eg PG P6, SGs might P6 P61 P62P63 P64 P65.. There is a GUI option to do this for both MOLREP & Phaser. You can guess the likely no of molecules using matthews In the GUI - that is in the MOL REp task - option analyse. Eleanor On 29 April 2014 10:21, Dom Bellini wrote: > Dear Sudarshan, > > I have the feeling the your R factor after Molrep is really high because > the program has failed to produce a correct/not-partial solution, which > could be due to many things, but usually the main problem is how good your > search model is. The difference in spacegroups is not a problem as in MR > the model will be rotated and translated inside the new cell. > > The number of molecules that you need to search for is a tricky question. > You need to make some educated guesses with help from analysis of, for > example, crystal solvent content (e.g., Matthews program from CCP4) and/or > self-rotation maps (Molrep or Polarrfn programs, also available from CCP4 > suite). > > If you have never before used molecular replacement, however, I would > advise that you get some help/guidance from some more experienced > colleague/friend in the Department, since you may encounter quite a few > more other problems that would perhaps be a bit too long to go through by > email. Some of these problems could also be probably self-solved by looking > at some tutorials on Molecular replacements. > > Best wishes, > > D > > > > > > > From: Sudarshan Murthy [mailto:sudarshan.murthy...@gmail.com] > Sent: 29 April 2014 05:11 > To: ccp4bb > Subject: [ccp4bb] Problem with Rfactor > > Dear All, > > I am very new to the field of crystallography, I have a few questions > which are very basic and getting input from this forum would help me a lot. > > Firstly I am trying to solve a structure using MR Molrep. In the result > the Rfactors are really high how do I reduce the same?? ..When I tried with > Phaser I didnt get any solution at all.. Will der be a problem with the > space group?? like the model is in monoclinic and the data which I have is > processed in hexagonal. > > Secondly while using hexagonal space group should I search for only one > monomer in Molrep?? > > These are very basic questions if anybody could help me out with this , I > would be very happy for the same. > > Sudarshan .N. Murthy > Crystallography Division > Bangalore >
Re: [ccp4bb] Problem with Rfactor
Dear Sudarshan, I have the feeling the your R factor after Molrep is really high because the program has failed to produce a correct/not-partial solution, which could be due to many things, but usually the main problem is how good your search model is. The difference in spacegroups is not a problem as in MR the model will be rotated and translated inside the new cell. The number of molecules that you need to search for is a tricky question. You need to make some educated guesses with help from analysis of, for example, crystal solvent content (e.g., Matthews program from CCP4) and/or self-rotation maps (Molrep or Polarrfn programs, also available from CCP4 suite). If you have never before used molecular replacement, however, I would advise that you get some help/guidance from some more experienced colleague/friend in the Department, since you may encounter quite a few more other problems that would perhaps be a bit too long to go through by email. Some of these problems could also be probably self-solved by looking at some tutorials on Molecular replacements. Best wishes, D From: Sudarshan Murthy [mailto:sudarshan.murthy...@gmail.com] Sent: 29 April 2014 05:11 To: ccp4bb Subject: [ccp4bb] Problem with Rfactor Dear All, I am very new to the field of crystallography, I have a few questions which are very basic and getting input from this forum would help me a lot. Firstly I am trying to solve a structure using MR Molrep. In the result the Rfactors are really high how do I reduce the same?? ..When I tried with Phaser I didnt get any solution at all.. Will der be a problem with the space group?? like the model is in monoclinic and the data which I have is processed in hexagonal. Secondly while using hexagonal space group should I search for only one monomer in Molrep?? These are very basic questions if anybody could help me out with this , I would be very happy for the same. Sudarshan .N. Murthy Crystallography Division Bangalore
[ccp4bb] Problem with Rfactor
Dear All, I am very new to the field of crystallography, I have a few questions which are very basic and getting input from this forum would help me a lot. Firstly I am trying to solve a structure using MR Molrep. In the result the Rfactors are really high how do I reduce the same?? ..When I tried with Phaser I didnt get any solution at all.. Will der be a problem with the space group?? like the model is in monoclinic and the data which I have is processed in hexagonal. Secondly while using hexagonal space group should I search for only one monomer in Molrep?? These are very basic questions if anybody could help me out with this , I would be very happy for the same. Sudarshan .N. Murthy Crystallography Division Bangalore