- IMP-users - salilab.org

localisation densities
by Andrea Graziadei 09 Jan '20

09 Jan '20

Hello, I am wondering how to extract localisation probability densities from a list of RMF files and density ranges? The "actin tutorial" code results in densities that are not in the same reference frame as the models, making it impossible to add the LPDs of the selected cluster to an ihm model, or to view them on top of the rmfs. Is there an example for this? Many thanks, Andrea

2 1

frame 1 score inf
by Atefeh Rafiei 08 Jan '20

08 Jan '20

Hi All, I am getting the following error and wondering what it is, what can cause it and if it is possible to bypass it. --- frame 1 score inf --- writing coordinates Traceback (most recent call last): File "MT-DCX-EM-6tub.py", line 281, in <module> mc1.execute_macro() File "C:\Program Files\IMP-2.10.1\python\IMP\pmi\macros.py", line 545, in execute_macro output.write_pdb_best_scoring(score) File "C:\Program Files\IMP-2.10.1\python\IMP\pmi\output.py", line 367, in write_pdb_best_scoring exec(open(self.best_score_file_name).read()) File "<string>", line 1, in <module> NameError: name 'inf' is not defined It is for modeling of a protein system that was working well when using only crosslinking data, but giving me this error when trying to use either EM or EM + XL as scoring functions. The same EM map and configuration is working fine for a very similar domain (same fold, very similar size). So not sure why it wouldn't work in this case. Regards, Atefeh

2 1

Error in Conjugate Gradient Optimization
by Altair Hernández Chinchilla 07 Jan '20

07 Jan '20

Hi there, I am currently using IMP to generate several models in which I apply a set of 62 restraints, plus ExcludeVolume (I work with XYZR sphere decorators). When I try to run the script 1000 times in order to generate 1000 different models, it work. However, when I try to run it 50.000 times, it came up with an error: *The script does the following:* o = IMP.core.ConjugateGradients(model_to_optimize) o.set_scoring_function(scoring_function) # this method return the scoring function that is being used o.optimize(num_iterations) # The maximum number of iteration of the optimizer to perform. # Returns the final score *When I try to run it a lot of times, the following error come up:* """ Traceback (most recent call last): File "/home/altair/bin/pycharm-community-2019.2.3/helpers/pydev/pydevd.py", line 1415, in _exec pydev_imports.execfile(file, globals, locals) # execute the script File "/home/altair/bin/pycharm-community-2019.2.3/helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfile exec(compile(contents+"\n", file, 'exec'), glob, loc) File "/home/altair/PycharmProjects/IMP/SCRIPTS_ALTAIR/scripts/exocyst_architecture.py", line 18, in <module> * n=50000*, tags=False, log_file=False) File "/home/altair/PycharmProjects/IMP/SCRIPTS_ALTAIR/scripts/exocyst_main.py", line 655, in create_models * conjugate_gradient_optimization*(model, sf, num_iterations) # OPTIMIZATION STEP # File "/home/altair/PycharmProjects/IMP/SCRIPTS_ALTAIR/scripts/exocyst_main.py", line 223, in conjugate_gradient_optimization * o.optimize(num_iterations)* # The maximum number of iteration of the optimizer to perform. File "/home/altair/anaconda3/lib/python3.7/site-packages/IMP/__init__.py", line 5804, in optimize return _IMP_kernel.Optimizer_optimize(self, max_steps) _IMP_kernel.ModelException: Failure in ConjugateGradients Process finished with exit code 1 """ What does it mean? Is it a problem on the max_steps of the Optimizer function? Why is this error comming up in the iteration 6000 ,e.g and not in the first one? Thanks you! Sincerely, Altair

2 3

IMP 2.12.0 release
by Ben Webb 06 Dec '19

06 Dec '19

IMP 2.12.0 is now available. This is contains major updates, most notably support for Python 3.8 and RHEL/CentOS 8, and a new IMP::bayesianem module for Bayesian handling of cryo-electron microscopy density map data. See https://integrativemodeling.org/2.12.0/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

mmcif creation with DNA
by Lenz, Swantje 05 Dec '19

05 Dec '19

Hello, I am building a system using PMI and IMP.pmi.topology.TopologyReader. I have proteins, DNA and RNA. In my PDBs, the DNA is encoded as DA, DC,DT,DG and the RNA as A,C,U,G. In topology, I specify the objects as DNA and RNA: |DNA1N |dark magenta|myfasta.fasta|DNAN,DNA |longRNAR1 |magenta |myfasta.fasta|longRNAR,RNA| If I have A,C,G,U and T in my fasta file I get the following warnings and errors for DNA.WARNING: Inconsistency between FASTA sequence and PDB sequence. FASTA type 1 "G" and PDB type "DG" WARNING: Inconsistency between FASTA sequence and PDB sequence. FASTA type 2 "G" and PDB type "DG" WARNING: Inconsistency between FASTA sequence and PDB sequence. FASTA type 3 "G" and PDB type "DG"The script then continues and modeling proceeds. However, when I try to implement the "deposition" part covered in https://integrativemodeling.org/tutorials/deposition/ , there are several errors due to not recognizing residue types, since the alphabet is defined as the peptide alphabet in the ihm module by default. How should the DNA be encoded in the fasta file? Additionally, we have two copies of one protein. This causes the second copy to not be created as an asymmetric unit in create_component in the protocol output of mmcif.py. We could fix this by changing the function as followed: def create_component(self, state, name, modeled, asym_name=None): if asym_name is None: asym_name = name new_comp = name not in self._all_components self._all_components[name] = None if modeled: state.all_modeled_components.append(name) self.asym_units[asym_name] = None # this was originally in the if statement below if new_comp: # assign asym once we get sequence self.all_modeled_components.append(name)? Thanks, Swantje ? -- ________________________________ Swantje Lenz M. Sc. Biotechnology Fachgebiet Bioanalytik (TIB 4/4-3), Institut für Biotechnologie, Technische Universität Berlin Technische Universität Berlin | Gustav-Meyer-Allee 25 | Gebäude 17a | 13355 Berlin Tel +49 30 314-72906 | web: http://www.bioanalytik.tu-berlin.de/

2 1

Maximum distance restraint between spheres
by Altair Hernández Chinchilla 04 Dec '19

04 Dec '19

Hi IMP users, I want to create pairwise distance restraints between IMP particles (spheres XYZR of a given radius) with a maximum distance condition.So I want a given sphere to be in a range between 0 and max distance of 17 Aº. How can I do it in IMP? Given a set of sphere pairs, and a distance between them, how can I apply a IMP function to force a maximum distance restraint in a similar way then a Harmonic function? f = IMP.core.Harmonic(float(tag_list[1]), float(tag_list[2])) s = IMP.core.DistancePairScore(f) r = IMP.core.PairRestraint(model, s, (bait_p, tag_p)) How can I substiute the IMP.core.Harmonic with a maximum distance function? Thanks in advance! Altair

2 2

Ambiguous restraints
by Francesco RAIMONDI 05 Nov '19

05 Nov '19

Dear all, I have a set of MS/XL restraints which I'd like to use to predict the structure of a putative homodimeric protein. Some of the XL are expected to happen within the same chain, others between the two distinct chains of the homodimer. The way I tackled the ambiguity issue so far is to explicitly generate all the possible XL/chain combinations and use this expanded list as an input for an adapted version of the human complement modeling script ( https://salilab.org/Complement) using IMP v2.9. Is there a better way to deal with ambiguity in IMP? I also found the AmbiguousCompositeRestraint <https://integrativemodeling.org/2.6.1/doc/ref/classIMP_1_1pmi_1_1restraints…> class, but didn't yet manage to find an example script using it. I'd be grateful for any suggestions you could provide me in this respect. Thanks so much and all the best, Francesco

2 2

Compiling a module out of tree
by Patrick Goetz 04 Nov '19

04 Nov '19

I'm trying to incorporate an out of tree IMP module called bayesianem. What I thought should be possible is to install the IMP Ubuntu 18.04 binary package, build bayesianem from source, and add it, but am running in to some problems. According to the IMP documentation (https://integrativemodeling.org/2.11.1/doc/manual/uselibrary.html) "When IMP is built or installed, it creates a CMake file called IMPConfig.cmake which contains information about how IMP was configured and where all the parts of IMP can be found." Unfortunately I could not find such a file anywhere on the system, but the bayesianem cmake script is clearly looking for it. Excerpted from bayesianem's source dir CMakeLists.txt: find_package(IMP REQUIRED) include(${IMP_USE_FILE}) and if I try and do the build I get: -------------------------------- By not providing "FindIMP.cmake" in CMAKE_MODULE_PATH this project has asked CMake to find a package configuration file provided by "IMP", but CMake did not find one. Could not find a package configuration file provided by "IMP" with any of the following names: IMPConfig.cmake imp-config.cmake -------------------------------- I also cloned the IMP git repo, but couldn't find anything there, either. The "Building modules out of tree" page (https://integrativemodeling.org/2.11.1/doc/manual/outoftree.html) talks about a FindIMP.cmake file, but I couldn't find this anywhere either. Presumably I am doing something wrong. Can someone give me some pointers? The bayesianem git repo (https://gitlab.pasteur.fr/rpellari/bayesianem) is pretty sparse on documentation. Thanks.

2 2

Compiling an IMP module out of tree
by Goetz, Patrick G 30 Oct '19

30 Oct '19

I'm trying to incorporate an out of tree IMP module called bayesianem. What I thought should be possible is to install the IMP Ubuntu 18.04 binary package, build bayesianem from source, and add it, but am running in to some problems. According to the IMP documentation (https://integrativemodeling.org/2.11.1/doc/manual/uselibrary.html) "When IMP is built or installed, it creates a CMake file called IMPConfig.cmake which contains information about how IMP was configured and where all the parts of IMP can be found." Unfortunately I could not find such a file anywhere on the system, but the bayesianem cmake script is clearly looking for it. Excerpted from bayesianem's source dir CMakeLists.txt: find_package(IMP REQUIRED) include(${IMP_USE_FILE}) and if I try and do the build I get: -------------------------------- By not providing "FindIMP.cmake" in CMAKE_MODULE_PATH this project has asked CMake to find a package configuration file provided by "IMP", but CMake did not find one. Could not find a package configuration file provided by "IMP" with any of the following names: IMPConfig.cmake imp-config.cmake -------------------------------- I also cloned the IMP git repo, but couldn't find anything there, either. The "Building modules out of tree" page (https://integrativemodeling.org/2.11.1/doc/manual/outoftree.html) talks about a FindIMP.cmake file, but I couldn't find this anywhere either. Presumably I am doing something wrong. Can someone give me some pointers? The bayesianem git repo (https://gitlab.pasteur.fr/rpellari/bayesianem) is pretty sparse on documentation. Thanks.

1 0

BuildModel 2 BuildSystem
by Marcell Miski 26 Sep '19

26 Sep '19

Dear Imp-Users, I am trying to model a complex using IMP however, I can't use the older IMP on a server. My problem is that when I try to change to using BuildSystem, my simulation fails. I attach, the original modeling.py and the new modeling.py with the two inputs. Can you help me how to fix this? Best regards, Marcell Miski Ph.D. Student PPCU FIT BP HU

2 4