MRBUMP (CCP4: Supported Program)


MRBUMP - automated search model generation and automated molecular replacement


mrbump hklin foo_in.mtz seqin foo.seq hklout foo_out.mtz xyzout foo.pdb
[Key-worded input file]


MrBUMP has three main parts:

  1. For a given target sequence, automated discovery of chains, domains and multimers that are possible templates for molecular replacement search models
  2. Preparation of actual search models using a variety of structure editing techniques
  3. Running molecular replacement using these search models and testing whether the resulting solutions will refine.

Note that MrBUMP makes a number of calls to web-based applications. If your sequence information is in any way sensitive, it is recommended that you use the option to run the fasta search locally rather than via the OCA web application. This will require installing fasta34 on the users local machine. The software can be downloaded from the EBI website here.


Before MrBUMP can be used, the following dependencies should be installed on the local system.

MrBUMP also requires that the local machine has a connection to the internet (directly or via a proxy).



Input structure factor file for target structure. Must include a FreeR_flag column. In general, the spacegroup in HKLIN is assumed to be correct (you should check this!). The only exception is the ENANt keyword which requests that both members of a pair of enantiomorphic spacegroups are checked by MrBUMP.


Input sequence file for the target structure. Can be in PIR or Fasta format or it can just contain the amino acid sequence.


MTZ file from Refmac5 refinement of the top MR solution.


PDB coordinate file from Refmac5 refinement of the top MR solution.


There are a number of options for specifying parameters (e.g. number of molecules expected in the asymmetric unit) or preferences (e.g. which multiple alignment program to use). Most of these have sensible defaults. The only mandatory keywords are LABIn and JOBId.

Compulsory keywords:


Main keywords:

ROOTdir, NMASu, MDLNum, MRNUm, ENSEmnum, INCLude, LOCAlfile, FIXEd_xyzin, IGNOre, MRPRograms, MAPRogram, MDLDpdbclp, MDLPlyala, MDLMolrep, MDLChainsaw, SSMSearch, SCOPsearch, PQSSearch, DOFAsta, FASTalocal, PACK, NCYC, UPDAte, ONLYmodels, TRYAll, USEAcorn, ACORnres, ENANt

Additional keywords:

CLUSter, CLEAn, DEBUg, PROXyserver

LABIN <program label>=<file label>...

Mandatory. This keyword tells the program which columns in the MTZ file should be used as native structure factors, sigmas, and FreeR flag. Available program labels are F, SIGF and FreeR_flag.

JOBID <job name>

Mandatory. This is a name for the job. A directory called "search_JOBID" will be created in the directory in which mrbump is started from. This directory will contain all of the downloaded files and results.

ROOTDIR <directory>

The root directory where the search folder will be created.
[Default Current working directory]

NMASU <number>

The number of molecules in the asymmetric unit. Leave this blank for automatic calculation.
[Default Automatic]

MDLNUM <number>

The number of template models to be prepared for molecular replacement.
[Default 50]

MRNUM <number>

The number of prepared models to be used molecular replacement.
[Default 20]

ENSEMNUM <number>

The number of prepared models to be used in a Phaser Ensemble.
[Default 5]

INCLUDE <pdb chain id 1> <pdb chain id 2>...

A list of PDB ID codes and Chain IDs to be included in the homologue search. Any specific chains entered here are automatically processed in molecular replacement regardless of how they score in the template model scoring.
Example: INCLUDE 1nio_A.

LOCALFILE <pdb filename [CHAIN chain id] >

Use this keyword to specify the location of a local PDB file to be used as a search model in mrbump. The full path to the file must be specified. The optional CHAIN subkeyword can be used to specify a particular chain in the PDB file. If CHAIN is not used then the program will extract chain "A" from the file.
Example: LOCALFILE /tmp/1nio.pdb CHAIN A.

FIXED_XYZIN <pdb filename IDEN sequence identity >

The FIXED_XYZIN keyword allows the user to input a fixed component structure in the MR search. This component should already have the correct orientation. This keyword can be used multiple times if there are more than one fixed components known. The sequence identity for each component against its corresponding section of the target sequence is required. The full path to the PDB file should also be specified. Fixed components are passed to both Phaser and Molrep.
Example: FIXED_XYZIN /tmp/fixed.pdb IDEN 0.43

IGNORE <pdb id 1> <pdb id 2>...

A list of PDB ID codes to be ignored in the homologue search. Used for development purposes.

MRPROGRAMS [ Molrep | Phaser ]

Names of Molecular Replacement programs to try search models in. Options are Molrep, Phaser or both. If both are selected Molrep will be used first.
[Default Molrep]


Name of the sequence alignment program to be used to do multiple alignment of the template structure sequences and the target sequence. In good cases, these programs should give the same result. In more marginal cases (e.g. small number of sequences, low sequence identity) they may give very different results.
[Default MAFFT]

MDLDPDBCLP [ True | False ]

If true models will be prepared for MR using the PDBclip method. With this method, the waters and hydrogens are removed from the coordinate file and the most probable side-chain confirmations are selected. If chain ID's are missing they are added.
[Default True]

MDLPLYALA [ True | False ]

If true Polyalanine models will be prepared for the MR step. All side-chains are removed from the PDB files.
[Default True]

MDLMOLREP [ True | False ]

If true models will be prepared using Molrep. Molrep does a sequence alignment of the target sequence and the template sequence and prunes the template structure file accordingly.
[Default True]

MDLCHAINSAW [ True | False ]

If true models will be prepared using Chainsaw. Chainsaw takes in a sequence alignment of the target sequence and the template sequence and prunes the template structure file accordingly.
[Default True]

SSMSEARCH [ True | False ]

If true mrbump will use the top match from the sequence-based search in a secondary structure-based search to find more potential homologues. Set to false by default. Requires perl and the perl SOAP-Lite module to be installed.
[Default False]

SCOPSEARCH [ True | False ]

If true mrbump will use the SCOP database to look for individual domains in the template structures found in the sequence-based and secondary structure-based searches.
[Default True]

PQSSEARCH [ True | False ]

If true mrbump will use the PQS service at the EBI to find more multimers based on the template structures found in the sequence-based and secondary structure-based searches.
[Default True]

DOFASTA [ True | False ]

If true, a FASTA search will be carried out to search for the possible template models. A user can turn this off and give specific chain IDs to be used as the template models. Note that at least one chain must be specified using the INCLUDE keyword if DOFASTA is set to False.
[Default True]

FASTALOCAL [ True | False ]

If true, the fasta sequence-based search will be carried out locally rather than via the OCA web-interface. This requires that the user have fasta34 installed on there system. This can be downloaded from the EBI site here.
[Default False]

PACK <number>

The number of clashes that Phaser will tolerate.
[Default 5]

NCYC <number>

The number of cycles of restrained refinement to use in Refmac.
[Default 30]

UPDATE [ True | False ]

If true, the search database files will be tested at the start of the job to see if they are out of date with respect to those available from the EBI website. If they are found to be out of date, the latest version will be downloaded.
[Default True]

ONLYMODELS [ True | False ]

If true, only the search models will be generated. The program will exit before any Molecular Replacement is carried out.
[Default False]

TRYALL [ True | False ]

If true, the program will try all of the search models in molecular replacement. If false the program will exit when it finds the first solution.
[Default False]

USEACORN [ True | False ]

If true, program will put each positioned and refined search model through the program Acorn to try and improve the phases. The target data must also be at least 1.7 A. Acorn is unlikely to help at lower resolutions, but this resolution limit can be changed with the ACORnres keyword.

MrBUMP prints out the correlation coefficient for medium E values from Acorn. An increase in these correlation coefficients over Acorn cycles is a good sign that you have the correct solution (the absolute value of the CC may be low, because these are not the strongest E values). The columns ECOUT, PHIOUT and WTOUT from Acorn can be used to generate high quality maps to help model re-building.
[Default False]

ACORNRES <resolution>

Resolution limit for applying the Acorn phase improvement procedure.
[Default 1.7]

ENANT [ True | False ]

If true, program will do molecular replacement for all search models in the enantiomorphic spacegroup, as well as in the HKLIN spacegroup, if an enantiomorph exists for the target data spacegroup. MrBUMP will identify the better spacegroup for each model. For good MR solutions, the correct spacegroup should be identified. For wrong or marginal solutions, it may be harder to distinguish the correct spacegroup.
[Default False]

CLUSTER [ True | False ]

If true, the model preparation and molecular replacement jobs will be farmed out to a cluster. Currently only works for Sun Grid Engine enabled clusters.
[Default False]

CLEAN [ True | False ]

If true, the program will remove the files generated for models that were marked as "Failed" solutions. Also, any files in the scratch area will be removed. This is to cut down on disk space usage.
[Default False]

DEBUG [ True | False ]

If true mrbump will give a more verbose output. Also, temporary directories will not be deleted at the end of the job.
[Default False]

PROXYSERVER <http_proxy server address>

If you need to use a proxy server to access the internet you should set it using PROXYSERVER. MrBUMP uses several on-line services and databases (e.g. the PDB) and thus requires internet access. It is possible to run MrBUMP without internet acces by turning off the FASTA, SSM and PQS searches and using locally stored PDB files as input search models. The proxy server is set in the environment in which MrBUMP is running. An example of a proxy server would be "".
[Default not set]


End keyworded input.


Simple example with minimal input using default values:

LABIN F=F SIGF=SIGF FreeR_flag=FreeR_flag

A more elaborate example:

IGNORE 1smw 1smm 1smu
DEBUG true
CLUSTER false 


Once a job has been started a user may view the current status of the job via the output log file or via the results.html web page which is created in the directory <ROOTDIR>/search_<JOBID>/results and is updated after each stage in the process. A set of search models is first generated and these are fed to the MR/refinement stage in sequence where the ordering depends on the alignment score of the template sequence against the target sequence. If a suitable solution is found, i.e. a model that refines well, the job will terminate and the final results will be displayed. The resulting refined PDB model and MTZ output from Refmac are made available to the user for further model building.


Ronan Keegan, Daresbury Laboratory, UK,
Martyn Winn, Daresbury Laboratory, UK


Norman Stein, Pryank Patel.

MrBUMP Program References

Any publication arising from use of MrBUMP should include the following reference:

R.M.Keegan and M.D.Winn (2007) Acta Cryst. D63, 447-457

In addition, authors of specific programs should be referenced where applicable:
Collaborative Computational Project, Number 4. (1994), "The CCP4 Suite: Programs for Protein Crystallography". Acta Cryst. D50, 760-763
W. R. Pearson and D. J. Lipman (1988), "Improved Tools for Biological Sequence Analysis", PNAS 85, 2444-2448
E.Krissinel and K.Henrick (2004), "Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions" Acta Cryst. D60, 2256-2268
A.G.Murzin, S.E.Brenner, T.Hubbard & C.Chothia (1995), J.Mol.Biol., 247, 536-540
K. Katoh, K. Kuma, H. Toh and T. Miyata (2005) "MAFFT version 5: improvement in accuracy of multiple sequence alignment" Nucleic Acids Res. 33, 511-518
Do, C.B., Mahabhashyam, M.S.P., Brudno, M., and Batzoglou, S. (2005) "PROBCONS: Probabilistic Consistency-based Multiple Sequence Alignment." Genome Research 15, 330-340
C.Notredame, D. Higgins, J. Heringa (2000) "T-Coffee: A novel method for multiple sequence alignments." Journal of Molecular Biology 302, 205-217
Chenna, Ramu, Sugawara, Hideaki, Koike,Tadashi, Lopez, Rodrigo, Gibson, Toby J, Higgins, Desmond G, Thompson, Julie D. (2003) "Multiple sequence alignment with the Clustal series of programs" Nucleic Acids Res 31, 3497-500
N.D.Stein (2006) in preparation
A.A.Vagin & A.Teplyakov (1997) J. Appl. Cryst. 30, 1022-1025
McCoy, A.J., Grosse-Kunstleve, R.W., Storoni, L.C. & Read, R.J. (2005). "Likelihood-enhanced fast translation functions" Acta Cryst D61, 458-464
G.N. Murshudov, A.A.Vagin and E.J.Dodson, (1997) "Refinement of Macromolecular Structures by the Maximum-Likelihood Method" Acta Cryst. D53, 240-255
E.Krissinel and K.Henrick (2005), "Detection of Protein Assemblies in Crystals", edited by M.R. Berthold, CompLife 2005, LNBI 3695, pp. 163-174. Springer-Verlag Berlin Heidelberg
Yao Jia-xing, Woolfson,M.M., Wilson,K.S. and Dodson,E.J. (2005) Acta. Cryst. D61, 1465-1475