MOSAICS-EM User Guide
Contents
1. topology file EM image header PDB file for the initial 3 point model refinement parameter file image parameter file defines multi scale natural moves at level 1 defines multi scale natural moves at level 2 defines multi scale natural moves at level 3 Segments of the Mm cpn are defined in the STRIDE record of the initial PDB file lidless 3pt pdb We can then group different segments into regions in the region files We subsequently represent the Mm cpn model using more numbers of smaller regions at hierarchical levels Fig 7 to describe finer conformational changes These levels are defined as follows Level 1 All the segments within the box are grouped into a single rigid f region in a way that chain breaks may occur between the stem loop and the equatorial domain The entire Mm cpn complex is treated as 16 rigid regions This level captures the overall rocking of the subunit while maintaining the Level 1 Level 2 Leyel 3 communication between adjacent Figure 7 Three levels of region compositions for a single subunits through the hand shake subunit with hierarchically increasing DOFs between the stem loop and NC termini Zhang et al 2010 It is defined in the file levell region data lel Level 2 In each Mm cpn subunit apical amp intermediate domain segments in one box belong to rigid region 1 The remaining segments in another box are grouped into another rigid region 2 Chain closures may o2. energy_scale 5 Below are the parameters that need to be modified for one particular image ea_az initial azimuthal angle for the model projection unit in degree ea_alt initial altitude angle for the model projection unit in degree ea_phi initial phi angle for the model projection unit in degree ea_range range to locally sample the around the current Euler angles unit in degree ea_interval interval for the local variation of the Euler angles az alt phi unit in degree pixel_size pixel size of the target image unit in A pixel resol_blur resolution to blur the model to match the target image unit in A expermnt_file path to the target image energy_scale weight for the EM energy In the above example we set ea_range 0 so no local optimization of the projection Euler angle is performed We can also introduce wrong initial Euler angle parameters and then let MOSAICS EM to refine the Euler angles as in the file orientation2 data cryo em parm pot typefnormal Nea az 0 ea_alt 8 ea_phi 0 lea rangef 2 Mea interval 1 In orientation2 data file we introduce an altitude deviation of 8 degrees We then let MOSAICS EM to optimize the Euler angles around the current ones between 2 degrees with an interval of 1 degree You can then run the refinement for both conformation and orientation by typing MOSAICS EM directory mosaics x refine euler input gt out If you have more than one
3. OPHEA 3 6 904 27 551 19 816 ATOM 9 CMA PHEA 3 4 541 24 650 17 440 eee eee ooo eee ooo eee eee eee Segment 2 Figure 3 An illustration of a model represented as 3 rigid segments connected by 2 flexible loops left and the first few lines of its corresponding initial PDB file with the STRIDE record defining segments and loops On the right hand side of Fig 3 shows the format of an initial PDB file corresponding to the model represented on the left The field CBLC A defines there is only one chain A If you have more than one chain such as Chain A and Chain B you can specify CBLC AB In the field STRIDE R means this residue belongs to a segment C means this residue belongs to a loop within which the chain closure needs to be solved To use the knowledge based potential each macromolecular residue is represented by a 3 point model that consists of the Ca carbonyl O atoms and a centroid CMA for the side chain If you have more than one chain you need to define each STRIDE for each chain Refinement parameter file refine input This parameter file is also used in the non EM version of MOSAICS which performs molecular simulation not related to the EM refinement Some parameters in this file are not related to the MOSAICS EM refinement but are still in this file for the completeness of the input For complete explanations of all the parameters of the refinement parameter file please refer to the MOSAICS user manual at http c
4. into the group II chaperonin chamber Cell 144 240 252 2 Minary P and Levitt M 2010 Conformational optimization with natural degrees of freedom a novel stochastic chain closure algorithm J Comput Biol 7 993 1010 3 Sim AYL Levitt M and Minary P 2012 Modeling and design by hierarchical natural moves Proc Natl Acad Sci U S A In press 4 Zhang J Baker M L Schroder G F Douglas N R Reissmann S Jakana J Dougherty M Fu C J Levitt M Ludtke S J et al 2010 Mechanism of folding chamber closure in a group II chaperonin Nature 463 379 383 5 Zhang J Ma B DiMaio F Douglas N R Joachimiak L A Baker D Frydman J Levitt M and Chiu W 2011 Cryo EM structure of a group II chaperonin in the prehydrolysis ATP bound state leading to lid closure Structure 9 633 639
5. the refinement is piped to a file called out A file with name sim_param out will be created in the current directory in which all the current refinement parameters will be recorded You can monitor the temperature of your Monte Carlo refinement by typing cat out grep Temperature You can monitor the acceptance ratio of your Monte Carlo refinement by typing cat out grep Chain 0 You can monitor the EM energy of your Monte Carlo refinement by typing cat out grep Cryo Input PDB file init pdb In order to use natural move DOFs a molecular model is represented as segments connected by flexible loops Rotational and translational degrees of freedom can be assigned to each segment and a chain closure algorithm Minary and Levitt 2010 is used to maintain chain connectivity and correct stereochemistry along the connected loop regions The cartoon on the left of Fig 3 is an illustration of how a molecular structure can be defined as several segments connected by flexible loops Segment 3 CBLC A STRIDE gt RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRCCCRRRRRRRRRR Segment 1 RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRCCCCRRRRRRRRRRRRRRRRRRRRRRRR RRRRRRRRRRRRRRRRR ATOM 1 CALS A i 8 857 20 435 16 184 ATOM 2 OLSA i 8 099 21 242 18 386 Loop 1 ATOM 3 CMALYSA 1 7 197 23 202 13 894 Loop 2 ATOM 4 CAVALA 2 8 945 24 024 18 951 ATOM 5 OVALA 2 8 872 25 910 17 477 ATOM 6 CMA VALA 2 10 414 24 061 19 450 ATOM 7 CAPHEA 3 6 047 26 111 18 105 ATOM 8
6. A This is overwritten in the region file if multi scale natural move is used Similar to Yprop tors sig but is used for chain closure Unit is in A number of refinement steps Output results every 100 refinement steps Unit of the output energy Ha atomic unit kcal kcal mol type of temperature modulation to use period for the temperature modulation amplitude for the temperature modulation baseline temperature for the temperature modulation random number to initialize the Monte Carlo The second section of refine input file sim_mol_def defines the basic parameters of the model and energy related to the MOSAICS EM refinement Target image file target hed and target img This is the target 2D image that you are refining against Class averages with high signal to noise ratio are usually used In this example we use some artificial data without any noise It is in the imagic format containing one header target hed and one actual image target img You can view it with any single particle EM image viewer such as the v2 command in EMAN Figure 4 the target image viewed with EMAN command v2 Image parameter file orientation data This is an example of the orientation data file which defines all the necessary parameters of the input target image scryo em parm pot typefnormal Nea az 0 ea_alt 0 ea_phi 0 lea range 0 Mea interval 5 pixel_size 2 resol_blur 10 expermnt _file target hed
7. Fig 2 Initial model target image deform model optimize amp projectio os seat match Figure 1 Refining model conformation and orientation MOSAICS EM EMAN2 1 image processing methods Figure 2 Architecture of MOSAICS EM MOSAICS 1 Conformational sampling methods 2 Energy minimization methods MOSAICS EM is capable of sampling the conformational space of the molecular model with much improved efficiency using natural moves at multiple scales We use Monte Carlo minimization with a modulated temperature profile to overcome local energy minima during optimization reference to our MOSAICS EM paper Installation of MOSAICS EM 1 EMAN2 is a software package developed by researchers at Baylor College of Medicine to perform single particle image processing for electron microscopy data To install EMAN2 you can use one of the two following options Option A Install EMAN2 by following the procedure at http blake bcm edu emanwiki EMAN2 Install Option B Precompiled EMAN2 libraries are provided in MOSAICS EM 2 Download MOSAICS EM 3 8 source code at http csb stanford edu minary mosaics download html 3 Untar version 3 8 EM tar gz and change the directory into source compile serial 4 Edit file Makefile and change the following two lines to your EMAN2 library and header directories INCLEM2 EMAN2 header files directory LIBEM2 EMAN library files directory EMAN2 header f
8. STANFORD UNIVERSITY MOSAICS EM User Guide Developers Junjie Zhang amp Peter Minary MOSAICS EM is a software package designed to refine molecular conformations directly against two dimensional 2D electron microscopy images By optimizing the orientation of the projection at the same time as the conformation it is particularly well suited to the 2D class averages from cryo electron microscopy By directly using projection images we relieve the urgent need for a density map that is not always available due to the structural heterogeneity or preferred orientations of the sample molecules on the grid Objective In our refinement procedure we change the atomic coordinates of the molecular model to increase its match to the electron microscopy EM projection images In addition we locally optimize the projection angle of the model to minimize the inaccuracy of the orientation parameters for the target image Fig 1 This is done iteratively with a Monte Carlo based optimization procedure We use Natural Moves to greatly reduce the degrees of freedom DOFs in the refinement Implementation of MOSAICS EM MOSAICS EM is built upon two software programs called MOSAICS Methodologies for Optimization and SAmpling In Computational Studies and EMAN2 Electron Microscopy ANalysis 2 We utilize the powerful sampling and minimization methods in MOSAICS while the basic image processing routines are called from EMAN library
9. _ def section region database file region data Please see the file refine region input You can then run the refinement with multi scale natural moves by typing MOSAICS EM directory mosaics x refine region input gt out This example shows how multi scale natural move can be used But little is gained by performing it on a small molecule such as the lysozyme In the next example we will demonstrate how multi scale natural moves can be used to greatly facilitate the refinement on a large macromolecular complex the Methonococcus maripaludis chaperonin or Mm cpn against a real experimental 2D cryo EM class average Using multi scale natural move to refine Mm cpn from the closed state to the open state with a single cryo EM class average Methonococcus maripaludis chaperonin or Mm cpn is a 16 subunit homo oligomeric chaperon from the mesophilic archaea It helps other proteins to fold in the archaea cell It consists of two back to back rings each containing eight subunits Each subunit has a substrate binding apical domain ATP binding intermediate and equatorial domains Mm cpn closes its folding chamber upon ATP hydrolysis and re opens after the y phosphate is released The entire complex is IMDa in size and the opening and closing of the ring is mostly achieved by a rigid body rocking of individual subunits The apical and intermediate domains are tightly coupled within a subunit by salt bridges at their domain inter
10. ccur between a the stem loop and the equatorial domain b the intermediate domain and the equatorial domain of the same subunit The entire Mm cpn complex contains 32 rigid regions In addition to the overall subunit rocking the relative motion between the equatorial domain and the other two domains are allowed It is defined in the file level2 region data Level 3 Based on Level 2 now we divide region 2 into 4 sub regions All sub regions have their own rotational and translational DOFs and they are kept connected by chain closures At this level more flexibility is introduced in the equatorial domain to describe more subtle conformational fluctuations around the ATP binding pocket It is defined in the file level3 region data You can run multi scale natural move refinement of Mm cpn at level 1 by typing cd levell MOSAICS EM directory mosaics x refine input gt out The model with the lowest EM energy at the current level is used as the initial model for the subsequent level The optimized Euler angles for that corresponding model at the current level are used as the initial Euler angles for the subsequent level We provide some useful scripts which can be downloaded from http csb stanford edu minary mosaics em scripts scripts tar gz References 1 Douglas N R Reissmann S Zhang J Chen B Jakana J Kumar R Chiu W and Frydman J 2011 Dual action of ATP hydrolysis couples lid closure to substrate release
11. face The communication between neighboring subunits within a ring is delivered by the B sheet that consists of the stem loop from one subunit and the NC termini from the other Douglas et al 2011 Zhang et al 2010 Zhang et al 2011 Based on this prior knowledge we defined the rigid segments and flexible connecting linkers as shown in Fig 6C c i iii API an n AR API INT gt i INT S S SL EQu 12 amp EQU T Cc NG Si Ss N ec lt q gt fa 5 8 fold axis Figure 6 A top view left and side view left of the lidless Mm cpn initial model in the closed state B top view 2D class average target image of the lidless Mm cpn in the open state C segments and connections as illustrated with three adjacent subunits Three subunits are labeled with I ii and iii API for apical INT for intermediate EQU for equatorial and SL for stem loop Here we use the lidless variant of Mm cpn so as not to deal with the unstructured region in the helical protrusion of the apical domains The example files can be downloaded from http csb stanford edu minary mosaics_em examples mmcpn tar gz Unzip file mmcpn zip you will get the following directories lidless 3pt pdb class average open 0 hed open 0 img levell refine input orientation data region data level2 region data levels region data pot database par 3pt prot na prm top database par 3pt prot na prm EM image file potential file
12. iles directory is where you put your EMAN2 header files EMAN2 Iibrary files directory is where you put your EMAN2 library file If you put the EMAN header files under EMAN2 include and EMAN library files under usr local lib then you specify INCLEM2 EMAN include LIBEM2 usr local lib 5 Type make and your C compiler will compile and make the executable file mosaics x in the directory called examples 6 Further installation instructions are available at http csb stanford edu minary mosaics install html Running MOSAICS EM with lysozyme artificial data To run a simple MOSAICS EM refinement the following files are required 1 init pdb input PDB coordinates of your model 2 refine input parameter file that defines global refinement parameters 3 target hed amp img input target EM image 4 orientation data parameter file for the target image 5 region data region file required if you want to use multi scale natural move DOFs 6 top_3pt_prot_na rtf topology file for the molecular model 7 par_3pt_prot_na prm potential energy file for the molecular model These files can be downloaded from link http csb stanford edu minary mosaics_em examples lysozyme tar gz Unzip this archive and change to its directory Run MOSAICS EM refinement by typing MOSAICS EM directory mosaics x refine input gt out MOSAICS EM directory is where the mosaics x file is The output information of
13. nd their corresponding definitions right in the region data file Fig 5 shows how the multi scale natural moves can be used by defining regions consisting of different segments Each region is assigned the independent degrees of freedom On the right hand side of Fig 5 are examples of the regions in the region data file with the parameters nseg number of segments in a region ncenter number of rotational center in a region segments_firstres the first residue for each segment segments_lastres the last residue in each segment segments_baseres the middle residue in each segment centers the residue used as the rotational centers for this region It can be either 1 or any of the residues defined in segments_baseres prop_trans_sig overwrite prop_trans_sig in refinement parameter file to define its value for each region prop_rot_sig overwrite prop_ rot sig in refinement parameter file to define its value for each region prop_trans_sig_freeres similar to prop_trans_sig but for each segments within a region unit in Set it to zero if no movement is allowed between each segment in a region prop_rot_sig freeres similar to prop_rot_sig but for each segments within a region unit in radians Set it to zero if no movement is allowed between each segments in a region The refinement parameter file also needs to be revised accordingly to use the region file One line is added in the sim_mol
14. sb stanford edu minary mosaics manual pdf Here we explain several parameters related to a particular MOSAICS EM refinement The first section of the refine input file sim_gen_def defines the necessary parameters to run the refinement Below are the basic parameters one may need to adjust for his own project using MOSAICS EM ssim gen defj simulation_typ MIN minimize_type stsamc prop_tors sig 0 prop_ rot sig 1 e 4 prop_ trans sig 1 e 3 prop_clos_sig l e 3 total_step_mc 7000 statistics freq 100 write_energy_unit Ha Istsamc type trigonom stsamc period 4000 stsamc_amp1 2500 stsamc_shift 0 random_seed 9378000501 the simulation type is minimization temperature modulated simulated annealing Monte Calo is used In each Monte Carlo step the newly sampled torsion angle between adjacent atoms in one segment is chosen from a normal distribution centered around the original angle with standard deviation o defined by prop_tors sig o The larger o is the broader the normal distribution and the higher the probability that a larger torsional step size is taken Unit is in radians Here we set it to 0 to make the segment rigid Similar to prop_tors sig but for the global rotation angles of a segment Unit is in radians This is overwritten in the region file if multi scale natural move is used Similar to prop_tors sig but for the global translation of a segment Unit is in
15. target images an image parameter file can contain multiple scryo em parm records with each one defines the parameters of its corresponding target 2D image This provides more experimental structural information since projections of more than one viewing angle are used But only use this option when the conformations captured by these images are identical Region file region data This representation was first introduced in the context of sampling by hierarchical natural moves Sim et al 2012 where the region elements were residues Here we further develop this technology to include segments as region elements and use it in our multi scale natural move refinement The multi scale natural moves are defined in this region file a region element_top_type segment r ee eu Janin Tn Fer gu dependency_type independent 4 nseg 2 Region ncenter 2 i segments_firstres A 1 A 89 segments_lastres A 39 A 129 segments_baseres A 20 A 109 centers A 20 A 109 prop_trans_sig 1 e 3 prop_rot_sig 1 e 4 prop_trans_sig_freeres 0 prop_rot_sig_freeres 0 region element_top_type segment dependency_type independent nseg 1 ncenter 1 segments_firstres A 43 segments_lastres A 84 segments_baseres A 63 centers A 63 prop_trans_sig 1 e 3 prop_rot_sig 1 e 4 prop_trans_sig_freeres 0 prop_rot_sig_freeres 0 Figure 5 An illustration of the customization of two regions left a
Download Pdf Manuals
Related Search