README

Here is an example describing the use of NestOR to compare six different coarse-grained representations of the Nucleosome Deacetylase (NuDe) sub-complex of the Nucleosome Remodeling and Deacetylase (NuRD) complex. The input and nude_modeling.py are adapted from the Integrative model of the NuRD subcomplexes repository.

NOTE: Prior to running the example, please download EM map from EMD22094, extract it, rename it as emd_22904.mrc and place it in input/gmm/ directory

Description of the example files

The examples/input comprises of the fasta sequences (examples/input/fasta), PDB structures (examples/input/pdb), crosslinking mass spectrometry data (examples/input/xlms) and an EM map along with the corresponding GMM representation (examples/input/gmm). Each target crosslink file from examples/input/xlms/original_xl_data is split into sampling_ and evicalc_ files as described in the main module README.

In addition it also comprises the topology{x}.txt files that define the representation to be used for running the nude_modeling.py modeling script. The {x} in the topology file's name corresponds to the number of amino acid residues to be coarse-grained to a single flexible bead for regions that lack a previously characterized structure. In this example, we are comparing the 1, 5, 10, 20, 30 and 50 residues per bead coarse-grained representations of the regions with unknown structure of NuDe sub-complex. It is to be noted that the regions with known structure will be modeled as 1 and 10 residues per bead representation. Importantly, note that any representation can be compared as long there is a valid modeling file associated with it (representation can be mentioned via a topology file or manually).

It also contains the nestor_params_optrep.yaml file which defines the NestOR parameters. Each parameter is described in the file itself.

The nude_modeling.py script is also adapted from the Integrative model of the NuRD subcomplexes repository for use with NestOR.

The nestor_output.yaml contains an example output for the given setup. In addition to this file, NestOR also saves a model from each iteration. These models are not included here due to space constraints. It also generates plots visualizing the log(evidence) (mean and standard error on the mean) (examples/trial_optrep_params_evidence_errorbarplot.png), MCMC per step sampling time (examples/trial_optrep_params_persteptime.png), NestOR total process time (examples/trial_optrep_params_proctime.png) and per step MCMC sampling time and log(evidence) together for all candidate representations (examples/sterr_evi_and_proctime.png).

Running nested sampling on this example

For this example, the user only needs to run ``` {path_to_IMP_installation} python nude_modeling.py `` This command will run one nested sampling run. However, for use with thewrapper_v{x}.pythe user will need to use modify thisnude_modeling.py<tt>a bit to use command-line arguments. In the current form, the command line arguments have been hard coded innude_modeling.py<tt>. The user will need to make thesesys.argv<tt>in the correct order for the wrapper to work. Please see the comments in thenude_modeling.py` for more details.