Integrative model of the NuRD subcomplexes

This repository is of the integrative model of the NuRD subcomplexes based on data from negative stain EM, chemical crosslinking, X-ray crystallography, DIA-MS, SEC-MALLS and COSMIC (Cancer Mutations Database). It contains input data, scripts for modeling and results including bead models and localization probability density maps. The modeling was performed using IMP (Integrative Modeling Platform).

These integrative structures were deposited in the PDB-Dev database with accession codes PDBDEV_00000152 (MHM), PDBDEV_00000154 (MHR), and PDBDEV_00000155 (NuDe). The negative-stain EM map used for modeling MHR was deposited in the EMDB database with accession code EMD-27557.

Directory structure

1. inputs : contains the subdirectories for the input data used for modeling all the subcomplexes.
2. scripts : contains all the scripts used for modeling and analysis of the models.
3. results : contains the models and the localization probability densities of the top cluster of the subcomplexes .
4. test : scripts for testing the sampling

Simulations

These are the independent simulations:

1. Modeling of MHR subcomplex : mhr
2. Modeling of MHM subcomplex : mhm
3. Modeling of NuDe subcomplex : nude
4. Modeling of MHR without using the EM data : mhr_xl_ctrl

Protocol

Sampling

To run the sampling, run modeling scripts like this \

for runid in `seq 1 NRUNS` ; do mpirun -np NCORES $IMP python scripts/sample/SUBCOMPLEXNAME_modeling.py prod $runid ; done

where,
$IMP is the setup script corresponding to the IMP installation directory (omit for binary installation),
SUBCOMPLEXNAME is mhr, mhm, nude, mhr_ctrl
NRUNS is the number of runs,
and NCORES is the number of cores on which replica exchange is to be carried out.

1. For MHR: SUBCOMPLEXNAME = mhr, NCORES = 8 and NRUNS = 50
2. For MHM: SUBCOMPLEXNAME = mhm, NCORES = 8 and NRUNS = 30
3. For NuDe: SUBCOMPLEXNAME = nude, NCORES = 8 and NRUNS = 50
4. For MHR without using the EM data: SUBCOMPLEXNAME = mhr_xl_ctrl, NCORES = 8 and NRUNS = 30

Analysis

1. Getting the good scoring models

Good scoring models were selected using pmi_analysis (Please refer to pmi_analysis tutorial for more detailed explaination) along with our variable_filter_v1.py script. These scripts are run as described below:

1. First, run run_analysis_trajectories.py as follows:
   $IMP run_analysis_trajectories.py modeling run_
   where, $IMP is the setup script corresponding to the IMP installation directory (omit for binary installation),
   modeling is the directory containing all the runs and
   run_ is the prefix for the names of individual run directories.
   Alternatively, one can also run the submit_run_analysis_trajectories.sh script from the scripts/analysis/pmi_analysis directory

2. Then run variable_filter_v1.py on the major cluster obtained as follows:
   $IMP variable_filter_v1.py -c N -g MODEL_ANALYSIS_DIR where, $IMP is the setup script corresponding to the IMP installation directory (omit for binary installation),
   N is the cluster number of the major cluster,
   MODEL_ANALYSIS_DIR is the location of the directory containing the selected_models*.csv.
   This can also be run using the submit_variable_filter_v1.sh script from the scripts/analysis/pmi_analysis directory.
   Please also refer to the comments in the variable_filter_v1.py for more details.

3. The selected good scoring models were then extracted using run_extract_good_scoring_models.py as follows:
   $IMP python run_extract_good_scoring_models.py modeling run_ CLUSTER_NUMBER
   where, $IMP is the setup script corresponding to the IMP installation directory (omit for binary installation),
   modeling is the path to the directory containing all the individual runs and
   CLUSTER_NUMBER is the number of the major cluster to be extracted.
   This can also be run using the script submit_run_extract_models.sh from the scripts/analysis/pmi_analysis directory.

2. Running the sampling exhaustiveness tests (Sampcon)

A separate directory named sampcon was created and a density.txt file was added to it. This file contains the details of the domains to be split for plotting the localisation probability densities. Finally, sampling exhaustiveness tests were performed using imp-sampcon as shown in scripts/analysis/pmi_analysis/*_sampcon.sh.
where, * is the name of the complex.

3. Analysing the major cluster

1. Crosslink violations were analyzed as follows:
   for xltype in adh bs3dss; do python get_xlink_viol_csv.py -c CLUSTER_NUMBER -m MODELANALYSIS_DIR -r modeling -k $xltype -t 35.0 & done
   and
   python get_xlink_viol_csv.py -c CLUSTER_NUMBER -m MODELANALYSIS_DIR -r modeling -k dmtmm -t 25.0
   One acn also use the get_xl_viol_validation_set.py script from the scripts/analysis/xlviol directory after changing the inputs section in the script as follows:
   $imp python get_xl_viol_validation_set.py -ia ../cluster.0.sample_A.txt -ib ../cluster.0.sample_B.txt -ra ../../MODEL_ANALYSIS_DIR/A_gsm_clust0.rmf3 -rb ../../MODEL_ANALYSIS_DIR/B_gsm_clust0.rmf3 -ta -ra ../../MODEL_ANALYSIS_DIR/A_gsm_clust0.txt -c ../cluster.0/cluster_center_model.rmf3 -x XL_FILE -t THRESHOLD
   where, XL_FILE is a file containing the crosslinks to be analysed.

2. The above scripts generate files mentioning the minimum distance for each crosslink. These files were then passed to xl_distance_hist_plotter.py as follows:
   python xl_distance_hist_plotter.py FILE_NAME XL_NAME THRESHOLD
   where, FILE_NAME is the name of the file,
   XL_NAME is the name of the linker used, and
   THRESHOLD is the distance threshold for that linker.
   This script will generate a histogram of the minimum distances spanned by the crosslinks.

3. Then, the files obtained from scripts in point 1 were passed to binner_cx-circos.pyas follows:
   python binner_cx-circos.py FILE_NAME
   where, FILE_NAME is the name of the file.
   This script generates a binned version of the input file which can then be used to make the crosslink plots using CIRCOS.

4. Contact maps were plotted for the NuDe models as follows: scripts/analysis/cosmic_and_distance-maps/submit_contact_maps_all_pairs_surface.py
   This script calls the scripts/analysis/cosmic_and_distance-maps/contact_maps_all_pairs_surface.py script. Please use --help for contact_maps_all_pairs_surface.py script for more details.

5. Finally, COSMIC cancer mutations were annotated on the models as follows:
   python color_mutations/color_mutation.py -i cluster.0/cluster_center_model.rmf3 -r 10 -mf mutations.txt

Results

For each of the simulations, the following files are in the results directory

• cluster_center_model.rmf3 : representative bead model of the major cluster
• chimera_densities.py : to view the localization densities (.mrc files)
• xl_violations.txt : list of violated crosslinks

For the NuDe models, mutation_colored_model.rmf and Distance_Maps are also added.

Information

Author(s): Shreyas Arvindekar, Matthew J. Jackman, Jason K.K. Low, Michael J. Landsberg, Joel P. Mackay, Shruthi Viswanath
Date: May 19th, 2022
License: CC BY-SA 4.0 This work is licensed under the Creative Commons Attribution-ShareAlike 4.0 International License.
Last known good IMP version:
Testable: Yes
Parallelizeable: Yes
Publications: Arvindekar, S, Jackman, MJ, Low, JKK, Landsberg, MJ, Mackay, JP, Viswanath, S. Molecular architecture of nucleosome remodeling and deacetylase sub-complexes by integrative structure determination. Protein Science. 2022; 31( 9):e4387. DOI: 10.1002/pro.4387.