IMP-users

imp-users@salilab.org

1 participants
273 discussions

Optimizing MCMC parameters in IMP
by Shruthi Viswanath 22 Nov '21

22 Nov '21

Hello IMP users and developers, Currently one usually tunes MCMC move sizes (e.g. rigid body maximum translation/maximum rotation, flexible bead maximum translation) and other sampling parameters (e.g. restraint weights, replica exchange maximum temperature) manually. Manual tuning can be a tedious process, not to mention anything about its accuracy. We introduce a method to automate the tuning of these sampling-related parameters called StOP (Stochastic Optimization of Parameters). Read more about the method here. <https://www.mdpi.com/2075-1729/11/11/1183> The software is available on github here <http://github.com/isblab/stop>. To get started, check out the tutorials here <https://github.com/isblab/stop/blob/main/tutorial/tutorial_basic.md> and here <https://github.com/isblab/stop/blob/main/tutorial/tutorial_replica.md> . We hope you will benefit from it. Feel free to email me or Satwik if you have any questions! Best, Shruthi

1 0

contact plots
by Andrea Graziadei 28 Sep '21

28 Sep '21

Hello, I have seen in some manuscripts (e.g. the cop9 signalosome paper by Gutierrez et al. https://www.pnas.org/content/117/8/4088 ) the analysis of average contact maps from clusters coming out of imp-sampcon or similar analysis of PMI modeling runs. In that manuscript for example it's in Fig. S7, panel G. I was wondering if you could tell me what is the tool used for this type of analysis, and whether it works on pdb or rmf files? How does it handle beads? I have used similar tools in the past, but those function only with PDB files, and for the system I am currently working with I am especially interested in contacts between coarse grained regions... Many thanks in advance, Andrea

3 2

REX score
by Altair Hernández 20 Sep '21

20 Sep '21

Hi, I'm running the Monte Carlo Replica Exchange sampling algorithm with IMP.pmi, which allows me to save the best N models. I'm interested in getting the rex score (IMP score) of the best N models after the sampling. So my question is: - Can the file *best.scores.rex.py <http://best.scores.rex.py>* be redirected to a desired folder from the function IMP.pmi.macros.ReplicaExchange0, I mean, before it is generated? The thing is that I'm running in parallel the same script, so I need to redirect this file to avoid it being replaced by another one. Or, is there any way to retrieve the IMP score from a ReplicaExchange object? Thanks in advance! Altair

2 1

IMP Silencing output
by Altair Hernández 20 Sep '21

20 Sep '21

Hi everyone, How can I silence warnings and output in IMP? I am using IMP.set_log_level(IMP.SILENT) but it is not working, and I have no clue about which argument I should put there (IMP.set_log_level(l = ?)). I would like it to work for both output (silencing verbose mode) and warnings. Thanks! Best, Altair

2 1

Paralellizing REX
by Altair Hernández 20 Sep '21

20 Sep '21

Hi, When sampling with IMP.pmi.macros.ReplicaExchange() in a cluster that uses slurm, an error comes out: [image: image.png] My question here is, can this function be parallelized with openmp (OMP) or/and with MPI? If so, which would be the more optimized combination between the two? Notice that in my case, I'm not running a long Replica Exchange simulation (with a high number of frames), but instead I'm iterating the whole process thousands of times with a few number of frames per modeling (50-100 frames/iteration). Finally, I'm interested in redirecting the scores output (best.scores.rex.py) to the same output folder that is created in the rex, since I want to rank all the final models according to the IMP score. Thanks in advance, Altair

2 1

extracting distance from CrossLinkingMassSpectrometryRestraint with state ambiguity
by Andrea Graziadei 16 Jul '21

16 Jul '21

Hello, I am modeling a multistate system in PMI using CrossLinkingMassSpectrometryRestraint in imp-2.15, and trying to adapt imp-sampcon for the purpose. I set up my system with 2 states having identical components with: topology = IMP.pmi.topology.TopologyReader(topology_file, fasta_dir=datadirectory, pdb_dir=datadirectory) topology2 = IMP.pmi.topology.TopologyReader(topology_file, fasta_dir=datadirectory, pdb_dir=datadirectory) bm = IMP.pmi.macros.BuildSystem(m) bm.add_state(topology) bm.add_state(topology2) root_hierarchy, dof = bm.execute_macro(max_rb_trans=rb_max_trans, max_rb_rot=rb_max_rot, max_bead_trans=bead_max_trans) and later set up crosslinking restraint (after creating crosslink database) by: cl_dsso = IMP.pmi.restraints.crosslinking.CrossLinkingMassSpectrometryRestraint(root_hier=root_hierarchy, database=cldb2, length=21.0, slope=0.01, resolution=1.0, filelabel='dsso', label="dsso", linker=ihm.cross_linkers.dsso, weight=1) After running replica exchange, i am trying to extract distances for my crosslinks (in theory 2 per protein copy, one in state 0 and one in state 1). However, when I for example extract distances from the stat file by stat_file=IMP.pmi.output.ProcessOutput('rmfs/0.rmf3') #extract distances for a specific crosslink restraint stat_file.get_fields(['CrossLinkingMassSpectrometryRestraint_Distance_|dsso|1278.1|Protein1|148|Protein2|689|1|PSI|']) I get a single distance per frame, corresponding to the distance in the latest state to be added, not to the shortest distance in the ensemble, and not one distance per state, as would make sense for analysing crosslink violations. As far as I can tell the score is behaving correctly overall (optimising for one state to be below the cutoff). I can still successfully run the first part of sampcon using score cutoffs rather than distances. But I wonder if it is possible to extract the shortest distance for each crosslink? Have I set up the multistate system incorrectly? (I am trying to work out how to set up the alignment etc for rmsd in such a case, but I have a long way to go there.) I can provide a test case if needed. Many thanks in advance, Andrea

2 2

Normal Mode Analysis
by Altair Hernández 11 Jun '21

11 Jun '21

Hi there, I am studying subunit mobility on a multimeric protein complex, after sampling with the Monte Carlo method. I would like to analyse the mobility of each subunit by using normal modes (normal mode analysis) of each subunit. Is there any function/method in IMP to do this? I'm using the PMI modeling protocol, working with a cryo-EM (coarse-grained), and iterating the sampling protocol 10.000 times (exploring from 10.000 different initial configurations). Thank you in advance! Altair

2 1

IMP 2.15.0 release
by Ben Webb 02 Jun '21

02 Jun '21

IMP 2.15.0 is now available. This is contains minor updates, most notably improved handling of MRC files, a new scoring function for cryo-EM data, and support for Apple Silicon. See https://integrativemodeling.org/2.15.0/doc/manual/changelog.html for the full change log. Check it out at https://integrativemodeling.org/ Ben -- ben(a)salilab.org https://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

Plotting XL restrains histograms from a symmetric complex on ChimeraX
by Gorbea, Jose 08 Mar '21

08 Mar '21

Hi all, I was wondering if somebody has experience plotting crosslink distance histograms for symmetric complexes. Specifically, since the complex I'm currently working with is symmetric, I obtain many copies of the same pseudobond for displaying on ChimeraX, each based on different atom sets. For example: Chain A, residue 160 - Chain A, residue 240; Chain A, residue 160 - Chain B, residue 240; Chain A, residue 160 - Chain C, residue 240; ... Chain G, residue 160 - Chain H, residue 240; Chain H, residue 160 - Chain H, residue 240; Some of these result in exceedingly long distances that are simply not correct. I know I can set an upper threshold through the "select maximumDistance xx" command, but I am looking for a way to obtain the set of shortest crosslinks for each residue pair combination in this assembly, even though some pairs may exceed an expected distance maximum (like a 30Å cutoff for a Ca-Ca distance) for other purposes. Any help with this would be much appreciated! Best, Jose Gorbea

2 1

ReplicaExchange error
by Altair Hernández 08 Mar '21

08 Mar '21

Hi, I am getting an error when running multiple integrative modelling processes with pmi, in some cases this error came out: [...] ------ simulated_annealing False ------ simulated_annealing_maximum_temperature 5.0 ------ simulated_annealing_maximum_temperature_nframes 20 ------ simulated_annealing_minimum_temperature 1.0 ------ simulated_annealing_minimum_temperature_nframes 100 ------ stat_file_name_suffix stat ------ write_initial_rmf True --- frame 0 score 850716.6597786957 --- writing coordinates *WARNING No frames were saved to file "RestraintSaveLink1" even though objects were added.WARNING No frames were saved to file "HierarchySaveLink1" even though objects were added.Traceback (most recent call last): * *File "modelling3_1.py", line 210, in <module> mc1.execute_macro() File "/aplic/noarch/software/imp/2.13.0-Python-3.7.4-GCCcore-8.3.0/lib64/python3.7/site-packages/IMP/pmi/macros.py", line 524, in execute_macro output.write_pdb_best_scoring(score) File "/aplic/noarch/software/imp/2.13.0-Python-3.7.4-GCCcore-8.3.0/lib64/python3.7/site-packages/IMP/pmi/output.py", line 500, in write_pdb_best_scoring exec(fh.read()) File "<string>", line 2 534.8805170496773]* In some other cases, all frames are correctly written to RMF but not the pdb_best_scoring models. I wonder if it is a simple lack of RAM memory, or lack of disk memory to write PDBs, or may be a combination of both things. Thanks in advance, Altair

2 1

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

IMP-users