ModelCraft

An automated model building pipeline for X-ray crystallography

Getting Started

ModelCraft can be installed using pip.

python3 -m pip install --user modelcraft

It also requires an installation of CCP4. The CCP4 environment needs to be set up for the pipeline to work. The simplest execution requires only a reflection data file in MTZ format and a sequence file in FASTA format.

modelcraft --hklin reflections.mtz --seqin sequence.fasta

Required Arguments

--hklin FILE
Input reflection data in MTZ format.

--seqin FILE
Input protein sequence in FASTA format.

Optional Arguments

--amplitudes COLS e.g. FP,SIGFP
Column labels for the observed amplitudes.

--convergence-cycles N default: 4
Number of cycles without improvement needed to stop the pipeline automatically. --convergence-tolerance is used to determine whether a cycle gave an improvement.

--convergence-tolerance default: 0.1
R-free difference needed to mark a cycle as an improvement. If the model at the end of a cycle has an R-free that is this much better than the previous best R-free then the cycle is marked as an improvement. --convergence-cycles states how many cycles without improvement are needed to stop automatically.

--cycles N default: 25
Maximum number of pipeline cycles.

--freerflag COL e.g. FreeR_flag
Column label for the free-R flag.

--help
Show an automatically generated help message.

--keep-jobs
Keep the files from intermediate jobs instead of deleting them. This can lead to large directory sizes.

--model FILE
Starting model. If input phases are not specified this will be refined using Sheetbend followed by REFMAC.

--no-auto-stop
Run the maximum number of cycles even if the model is not improving.

--phases COLS e.g. PHIB,FOM or HLA,HLB,HLC,HLD
Column labels for input phases as either a phase and figure of merit or Hendrickson-Lattman coefficients.

--remove-non-protein Use this option to remove non-protein residues from the starting model. By default, residues that are not protein or water (i.e. do not have a CA atom and consist of more than a single oxygen atom) will be kept and Buccaneer will avoid building within 2Å of them.

--semet
Build selenomethionine instead of methionine.

--twinned
Turn on twinned refinement. Only do this if you are sure your data are twinned.

--unbiased
Pass input phases to REFMAC for MLHL refinement.

Developer Arguments

--buccaneer FILE
Path to an alternative buccaneer binary.

Links

References

ModelCraft
Publication in progress
Buccaneer
K Cowtan. Acta Cryst. D62 1002-1011 (2006) DOI
REFMAC
O Kovalevskiy, RA Nicholls, F Long, A Carlon, GN Murshudov. Acta Cryst. D74 215-227 (2018) DOI
Coot
P Emsley, B Lohkamp, WG Scott, K Cowtan. Acta Cryst. D66 486-501 (2010) DOI
Sheetbend
K Cowtan, J Agirre. Acta Cryst. D74 125-131 (2018) DOI
Parrot
K Cowtan. Acta Cryst. D66 470-478 (2010) DOI