uDock 2.5 User Manual
Contents
1. uDock 2 5 Page 10 of 18 www shadnia com running on 64 clients number and list of processing nodes compute 1 15 compu ute 1 12 local running the remote SVL loader loading program on processing node V OK loading udock svl i Docker Clientgvl on Clienti seas V OK loading udeck svl i Toolbox syl on Clients ssis Loading Codes on all Clients successful gt gt Scanning the input database entry 1 of 5 first ligand conformer Training mode ON using 10000kcal mol limit for energies c try 0 to make a good initial pose max3000 generate the ligand inside the box c POUCELING ZX yu aaa a c O DEII 5 265 064 c discarding srira did mot fit lun the Dox c try 1 to make a good initial pose max3000 c POCACING a Salmus c CONLIN papapa c good pose found did fit in the Dox compute 1 29 local gt training total 1 s compute 1l 18 l cal gt training total 9 i comoute 0 18 local gt training total 10 Activating AutoPilot calculating tight energy thresholds gt updating the threshold for task 7 idx 3 value 100000 to 21613 4 gt updating the threshold for task 10 idx 3 value 100000 to 19 2877 gt updating the threshold for task 13 idx 3 value 100000 to 1 13228 training finished killing all current tasks start over with tight thresholds gt compute 1 13 local gt good pose total 1 good results to be saved in output gt compute 0 3 local2. pose meaning that the rotatable side chains of the active site residues need to be actively rotated The current version of uDock does not perform this but the program is designed to allow future implementation of this feature For many simple active sites sometimes called rigid active sites as those of nuclear receptors this 1s not crucial uDock 2 5 Page 4 of 18 www shadnia com viii The raw crystal coordinates need to be treated properly to generate an energy minimized structure that is as faithful as possible to the raw coordinates Three issues are remarkable in this topic 1 A regular energy minimization on raw crystal geometry causes unacceptable amount of perturbation in the geometry esp that of buried active sites For this reason the PDB Thawing program H PDB THAW see www shadnia com H_Thaw should be used to energy minimize the initial PDB file 2 The limited precision of Cartesian system causes small random perturbations of the macromolecular geometry independent of ligand binding For this reason the energy read outs in uDock are by default limited to the active site residues the 1 shell or S1 which are kept fully flexible The amino acid residues surrounding the active site labeled as the 2 shell or S2 are kept semi flexible and their energies are excluded from the final energy calculations The Atoms or residues surrounding S2 are kept rigid and can be deleted to accelerate the calculations The PDB
3. uDock 2 5 flexible docking program for MOE smp clusters User Manual H Shadnia Sep 2009 Contents Part A 1 Introduction 2 Theory Part B 1 Program configuration 2 Preparing the target structure 3 Preparing the docking box 4 Preparing the ligand database 5 Job setup and run 6 Post processing Part C Appendix I Preparation Flowchart Appendix II Docking Flowchart Part D Selected uDock citations Part E Tutorial A how to render the active site cavity Tutorial B how to generate a database of ligand conformers Tutorial C Docking of estrogenic ligands into ERa PART F Datasets PART G FAQ uDock 2 5 Page 2 of 18 www shadnia com Part A Introduction uDock is a flexible ligand receptor docking program designed for MOE www compchem com This program is available for download from www shadnia com The main idea behind uDock is to reduce the number of commonly used and potentially invalid approximations and cosmetic procedures In return the program requires considerable processing power The main workflow in uDock is shown in appendices I II In order to run a docking job three pieces of data are required a The structure of receptor This is prepared from the raw crystal coordinates using the PDB Thaw program and is saved as a moe file see section 2 b The coordinates of the docking box This 1s prepared using the crystal coordinates and the DockingBox program It is saved as a mdb fi
4. Thawing program defines these shells but they can be manually modified too The definition of these shells are saved within the receptor file that is generated by PDB Thaw 3 When comparing different instances of raw crystal structures of supposedly identical complexes conformational difference is observed which leads to variation in docking results A calibration test can be used to identify and filter out the 1ll modeled residues in each instance of crystal structure This results in dramatic improvement of the docking results H Shadnia J S Wright to be published uDock 2 5 Page 5 of 18 www shadnia com PART B 1 Program configuration Purpose installation and configuration of MOE and uDock l a Program files The following is the description of the individual programs Filename description oD eeka ev launcher configure and launch the docking jobs Ht DOCKSt u Host program He Dogker Clientesyl Client program H DockingBox sv1 Graphical interface to define the docking box H Host LOO Le ceva Host utilities H Toolbox svl Main utilities H PDB Thaw svl Graphical PDB Thaw program H DB Prepare svi Graphical program to prepare tag the input database H DB PickMin utility to pick the lowest energy conformers H DB PicknMin utility to pick n lowest eneroy conformers He POST PROCESS post processing tools H dpi SVI Program to analyze interaction forces and energies l b Setup the environment variable For b
5. gt good pose total 2 gt compute 0 18 local gt good pose total 3 compute 1 30 local gt bad pose total 7 bad results will be discarded ecompute 1 5 local gt good pose total 39 gt compute 0 13 local gt good pose total 40 x finished entry 1 in 197 secs estimated remaining 788 secs elapsed estimated time good 40 39 2 bad 62 60 8 total 102 overall statistics gt gt Scanning the input database entry 2 of 5 staring up with the next ligand DONE finished scanning the input database the last ligand is sent to the cluster gt scanning the input db finished job spool done Search finished Aborting further calcs on entry 6 Computations complete all calculations are done total calculation time 1172 0 seconds total time good 200 30 9 bad 448 69 1 wasted 0 total 648 performance stats The second file contains MOE s OS level messages and can be used for troubleshooting of MOE smp network This file would be named moe 5285 log for the previous example 6 Post processing 6 a Understanding the output database To allow for method development uDock saves a raw database of the docked poses The database contains the following fields mseq ligand number same as the input database uDock 2 5 Page 11 of 18 www shadnia com Conf conformer number str l receptor coordinates at the complex configuration str 2 ligand coordinates at the complex configurat
6. used to abort or accept a given pose can be easily reconfigured for a wide range of applications For most purposes however the dynamical Autopilot system with the default values is efficient uDock emphasizes in achieving convergence for each complex uDock 2 5 Page 3 of 18 www shadnia com The docking engine produces a raw output database which includes many conformations of complexes of ligand target with detailed energetics recorded uDock does not perform any cosmetic cleanup on this database A set of post processing programs then help to find the best complex and read the energetics and perform quantitative calibration or prediction of activities Section 6 and appendix 111 Theory After running extensive tests on different targets the following rules were used to design uDock 1 The initial energy of a randomly generated complex has NO meaningful correlation to the final energy minimized value This is simply because even one single VdW clash between two hydrogens can raise the energy to hundreds of thousands of kcals mol But after a few iterations of energy minimization these clashes will relieve and the energy becomes more stable and comparable to the final geometry il The receptor active site can not be treated as rigid due to the same reason iii Explicit water molecules have to be maintained and treated properly esp those which make a water bridge between two functional groups In gener
7. al they need to be re arranged or re docked for each individual complex uDock requires the explicit waters to be present but it doesn t re dock them by default iv Due to the gearing effect ref number of possible conformations or poses for the simplest ligand receptor or even host guest complexes can easy exceed thousands For this reason uDock tries at least 2 5000 poses for each ligand This is crucial to ensure reaching convergence 1 e subsequent runs of the program will not result in a lower energy complex Some programs apply smoothing functions to VdW terms in order to bypass this effect Such measures artificially reduce the complexity of conformational space but they often diminish the uniqueness of the active ligands which perfectly fit the grooves of the target active site vi Each conformer of the ligand is a unique chemical entity which may or may not fit to a target For this reason all room temperature conformers of each ligand need to be individually docked into the active site In rare cases where a particular moiety of the receptor is extraordinarily complex even different rotamers need to be tried When docking macromolecular structures as peptides since it isn t possible to generate all plausible conformers a reasonable number of such conformations generated by MD may be used vil In theory the entire conformational space of the target protein need to be explored for each individual
8. ases for example when a small molecule as phenol is compared to a larger molecule this term cannot be ignored Our group in collaboration with CCG is developing a method to estimate this term Oct 2009 6 c The detailed analysis of interaction energies We developed a program to analyze the interaction energy of individual amino acids with the active site residues The program calculates and visualizes the interaction energies as well as forces which allow for detailed analysis and understanding of molecular events for a given complex See www shadnia com H_IFE for more details and download HOH _8 0 306 The analysis of the interaction energies and forces The red arrows show the atomic forces The favorable interactions are shown in dark green 0 to 4 kcal mol and light green stronger than 4 kcal mol The magnitudes of interactions are displayed as labels We also prepared a comprehensive set of programs also allows automated comparison and analysis of such forces and energies for an entire set of complexes generated via docking Please email the author to obtain this particular package uDock 2 5 Page 13 of 18 www shadnia com PART C Appendix I Preparation flowchart Systematic Stochastic Conformational Analysis Raw PDB structure pasta apse individual databases Y of ligands Add hydrogens Rotamer generation Y Clean up Y PDB Thaw Y _complex moe gt Define the docking box Merge to one database Y remov
9. ash edit the file bashrc and add the following lines uDock udock svl1 export uDock For other types of shells refer to the operating system manual l c MOE should be installed and configured to run in SMP mode Please refer to MOE manual on Installing and Running MOE smp optional configure the grid engine load balancing system 1 d 1 d I Create a sun grid engine file name moe sge that contains the job creation details bin bash it CHOOSE THE NUMBER OF PROCESSOR WHEN YOU RUN THE JOB FOR EXAMPLE TO USE 8 PROCESSORS TYPE if qsub pe moe 8 moe sge it it uDock 2 5 Page 6 of 18 www shadnia com Tell Gridengine what shell to use S bin bash it F Star this script from the current working directory cwd Combine output and error messages into one output file to ol yY it Name of the output file ie O MOS a TOD TI ser echo n 1 am running T echo SINPUT opt moe bin moebatch mpu STMPDIR moemachines mpulog moe JOB ID log run S INPUT 1 d II create a job generation script named runMoe as follows bin bash usage runMoe numberOfProcessors inputFile qsub v NS 1 INPUT S2 pe moe 1 opt moe bin moe sge 2 Preparing the target structure Purpose to prepare an energy minimized model of the full or truncated macromolecular target 2 a Provide the 3D coordinates Download the coordinates file from the protein databank RCSB ORG or use the
10. e the co crystal ligand Prepare tag the database Y Y _receptor moe _dockingbox mdb _ligands mdb x uDock 2 5 Page 14 of 18 www shadnia com woo eIUpeYys MMM Q JO I seg 7 yooqn iSNOQ A SIU PeOsip SIU PAeOsIp SIU PAeOSIp SIU PeOsIp Sank A A A A A s s s s OU lt uop spuebi e A A A 4 A ou pl lt a ou lt l lt lg ou Zl lt u4 ou lt lt 1lu qd s aa pue i ae sosod seibioue y pees s 1 u au peed s I Bi u ou peo seilbioue ou peed jo saquuinu uBnou A u A A A Bulxl pue s194 9 jddy Buixl pue s1949 Ajddy Bulxi pue s1949 Ajddy Bulxi pue s194 9 ddv soibieue pue A A A A s im onus Y ALS puebl y 2919S puebl y 2919S puebl y 2919S puesi y 2919S e peas suolle19 Op SUOHeJ9 oZ suoneJ i OZ u P A6J U y ZIWIUIN ABJ u OU zIWIUIN A6J U OU ZIWIUIN E us E 91S Anoe ON SOA JOYS JSIY y 99l9S pouu BuluIe1 1 s amp X0q y SPISU SJ OU A uOol e SUeJ uonelu luo wopues e ayes9ueb puebl ay peol A xoq on peo 10 d 9 1 y peol peYyoOMoO SUTYOOG lI xipuoddv PART D Selected uDock citations Deconstructing the estradiol ABCD ring structure A new family of A CD compounds which are potent and selective estrogen receptor agonists Link Tony Durst Mohammed Asim James S Wright Hooman Shadnia Christi
11. er name This is the main target file that will be used for docking uDock 2 5 Page 8 of 18 www shadnia com 3 Preparing the docking box Purpose to define a virtual box that contains the active site random conformers of the ligands will be generated inside this box during the docking 3 a Load the thawed protein file _thawed moe 3 b Render the active site boundaries See tutorial A for an example 3 c Run the docking box program H DockingBox svl Modify the dimensions if necessary and save the final box See below ID x File Edit Selection Render Compute GizMOE Window Help SVL SEQ Cancel deed Builder Minimize Delete 7 Docking Box jol gt x save Filename _dockingbo mdb Origin X L T L L L L TT TTT 2 065 yT I T L L L TTTI 4o08 iu aa a Extent x ce eae yf ir Rotate C O ji S r Select P Spacing 0 375 CAUS Ligx Atom Monitor off inside box Residue Monitor off inside box AutoBox Padding 3 m Selected Atoms Only Rotate x TTHv O0 z D TIl Transtate x ITTHv Oz C0 a TI 4 Preparing the ligand database Purpose to prepare a tagged database of all conformers rotamers of all ligands to be docked into the receptor 4 a build conformers of each ligand 3 a I Use the molecular builder module or other ways to draw import the structure of the ligand 3 a II Explicit list of conformers rota
12. grid engine 5 d I To run the program on 24 processors use the following command runMoe 24 H Dock svi 5 d II To monitor the job use qstat The following shows an example of the docking job running on 24 processors The job has been assigned a job ID of 5285 It can be killed using qdel 5285 Job ID prior name user state submit start at queue slots ja task ID 5285 0 55500 moe sge hpc1764 09 11 2009 13 23 48 all q compute 0 21 local 24 5 d III To view the docking log When docking is running it creates two files which are named by job ID as explained above the error file contains all debugging messages that would be visible on MOE s SVL window in graphical mode This file would be named moe 5285 err for the previous example You can monitor the progress of docking by viewing the contents of this file using cat or more commands or any other method The following is an example comments are added in italics running uDock using the following options input parameters program path udock Ssvl proj ect path udock Lig dbname input mdb output dbname output mdb target fil ename receptor pha mos docking box filename dockingbox ERa mdb target good poses 40 target bad poses 3000 lig Max rerraes 2000 anto Pirlot ON Training 1berelions lt 10 Threshold Percent 60 _forceri ld SMOE lib mmrto4s tr file transfer adapted to OS unix automatically assigned
13. homology modeling to build it 2 b Add hydrogens In case the hydrogen atoms are missing add them using MOE gt Edit gt Hydrogens gt Add Hydrogens or use MOE gt Compute gt Protonate 3D 2 c Clean up the structure 3 c I Inspect atom types After selecting the appropriate force field MOE gt Window gt Potential Setup gt Force Field check the for missing parameters to ensure there are no illegal atom types or undefined force field parameters To do this use MOE gt Window gt Potential Setup gt Parameters check Display Missing Parameters Only and under Parameter see each individual parameter type most importantly Atom Type Bond Stretch Angle Bend Fig 1 uDock 2 5 Page 7 of 18 www shadnia com 7 Potential Setup MMFF94x j loj x Forcefield Parameters Restraints Wall Stretch Bend m Display Missing Parameters Only aO ec 2 c II Inspect missing atoms Load the sequence editor Main menu gt SEQ Enable synchronized selection Sequence Editor gt Selection gt Synchronize Load the Residue Selector Sequence Editor gt Selection gt Residue Selector Under More select Amino Acid Atom Data Click on each of four classes of missing data to find the missing atoms see below Correct the missing atoms if any 4 Residue Selector Ioj x Residues All Chains v All Clear Residues with Visible Atoms Invert Nonempty Residues Named Set Create Unio
14. ion Thus the most important energies are the ones at the end of the energy minimization those of step 4 and 5 R 4 sel The internal energy of the ligand at the bound conformation Er R 4 int The interaction energy of the ligand vs target Eint R 5 sel The internal energy of the pseudo complex See below for details 6 b Quantitative prediction To allow for fundamental analysis of the results the raw system energies are recorded in uDock This is done in 5 steps step 1 before any energy minimization step 2 after 20 iterations of energy minimization step 2 after 20 additional iterations step 3 after 40 more iterations final energies Steps 4 5 at the end of energy minimization Each energy reading generates three sets of numbers designated by suffices all sel and int fields which correspond to the total system energy the internal energy of the selected atoms and the interaction energy of the selected atoms vs the rest of the system Through step 1 to 4 the ligand structure is selected and at step 5 the ligand plus the first shell of amino acids active site are selected By definition the interaction energy is the difference in energy of products complex from those of the reactants ligand and receptor in the conformation they hold upon binding Ein Ec Er Er 1 In comparison the binding energy AEbinding is the difference in energy of products complex from those of the free ligand and receptor AEbindi
15. le See section 3 c The ligand database This includes all plausible conformers and sometime even rotamers of the ligands to be docked into the active site It can be prepared in different ways and is saved as a mdb file It also has to be properly prepared or tagged using H DB _ Prepare program see section 4 After the three files were prepared a job launcher job named H_Dock svl is edited to reflect the correct filenames file paths and some other user settings Then the docking will be run on a single machine or a windows Linux cluster The full flowchart of docking is shown in Appendix II Briefly random conformations of the each ligand database entry individual conformers are placed inside the docking box the geometry is optimized and the energies are recorded before during and at the end of energy minimization This is done for a user defined number of time usually 10 20 times which is named the training phase After that the program finds typical and minimum energy values for the given complex and tightens the acceptable energy thresholds for the rest of calculations on the given ligand conformer The energy calculations on ligand poses which are too high energy will be aborted to save the processing time The program will keep generating a user defined number of good solutions for each ligand conformer and proceeds to the next entry when done The energy minimization steps and the energy criteria
16. mers for each ligand should be prepared Use the systematic conformer search stochastic conformer search or even MD simulation to build an MDB containing the conformers for each ligand See tutorial B for an example 4 b merge multiple ligands into one database After preparing individual conformers use the database merge tool Database Viewer gt File gt Merge to merge all individual files into one large database 4 c prepare or tag the database Use the H DB Prepare svl program to tag the large database Tagging assigns ligand numbers specified in the mseq column and conformer numbers specified in conf column to the uDock 2 5 Page 9 of 18 www shadnia com database uDock uses these tags to monitor different procedures An example of prepared and tagged input database is shown here ay Database Viewer udock input mdb compound compound compound compound compound compound compound i at dy mo mo eo eo eo ae yl m Ble eo cn Bl co o 1 1 1 1 1 2 compound 2 2 2 2 S5 Job setup and run 5 a Open the launch program H Dock svl using the text editor or any other text editor 5 b edit the options to reflect the correct filenames and desired options Save the svl file 5 c run the launcher program in MOE smp mode See MOE s manual on Installing and Running MOE smp 5 d how to run the launcher program for HPCVL
17. n Extend Radius 4 J Y Froximity Pennortir fC railiiFn ri mr arii Striictira Low High f More Amino Acid Atom Data v Missing Atoms Incomplete Backbone Monstandard Atom Names Select Atoms 2 c II Delete unrelated chains Many PDB files include crystallographic or biological dimmers trimers or polymers They Atom Type ijk gt NF CH 43 179817 Sont Oa 43 173917 gt C C Angle Bend 43 179817 gt C C 43 179817 gt c fc tetch Bend 43 179817 HE E OutotFlane 43 179817 gt C Cc 43 173817 gt C c Torsion 43 17381 gt C Cc 43 173817 gt C C 499 C56 C 43 173817 gt C C 488 C56 C 43 173817 gt C e 5413 C ae 43 173817 gt C C 522 C56 C 43 179817 gt C C 522 C5B C 43 173817 gt C c 565 cC cc 43 173817 gt C C 568 C5B C 43 173817 gt C C 568 C5B C 43 173817 gt C C 647 C C 43 173817 gt C C 650 C56 C 43 173817 gt C 650 C56 C 43 179817 gt C c 635 C te 43 173817 gt C C 698 c56 C 43 173817 F oren Report dese might include co crystallized small organic molecules or metal ions Delete all unrelated structures 2 d run the Thawing Select the ligand on the screen and run the H PDB Thaw svl Select the appropriate protocol program and click OK Refer to http www shadnia com H_ Thaw for more information 3 e open the thawed file thawed moe and delete the ligand Save this file as _receptor moe or any oth
18. ne Pratt John Katzenellenbogen Kathryn Carlsen Bioorganic and Medicinal Biology 2009 Glu353 a Leu384 His524 B lle373 A i Investigation of Residues K112 E136 H138 G247 Y248 and D249 in the Active Site of YeastCystathionine B Synthase Link Pratik Lodha Hooman Shadnia Colleen M Helferty James Wright and Susan M Aitken Biochemistry and Cell Biology 2009 Design synthesis and biological evaluation of substituted 2 alkylthio 1 5 diarylimidazoles as selectiveCOX 2 inhibitors Link Latifeh Navidpour Hooman Shadnia Hamed Shafaroodi Mohsen Amini Ahmad Reza Dehpourd and Abbas Shafiee Bioorganic amp Medicinal Chemistry 15 1976 1982 2007 Phed18 lle517 Protein Targeting with Small Molecules Chemical Biology Techniques and Applications Hiroyuki Osada Wiley ISBN 978 0 470 12053 8 Hardcover WILEY uDock 2 5 Page 17 of 18 www shadnia com PART E Tutorials To be added soon Tutorial A how to render the active site cavity Tutorial B how to generate a database of ligand conformers Tutorial C Docking of estrogenic ligands into ERa PART F Datasets To be added soon PART G FAQ To be added soon uDock 2 5 Page 18 of 18 www shadnia com
19. ng Ec Er Ez 2 The free ligand and receptor L R are deformed upon binding to form the deformed ligand and receptor L R thus the deformation energies AE AEpr are defined as AEL Er Er 3 AErR En Er 4 Thus from eq 1 Ec Ein En Er From eq 3 E Ep AE From eq 4 Er Ep AER Now using eq 2 AE binding Em Ens Ey Er AER Er AB Binet Eee Eye Ege AEr Ep r Ar Ein AER AEr Thus three terms from the docking describe the binding energy l Eint is the ligand interaction energy designated as R 4 int in the raw database 2 AER The receptor deformation energy since the energy of free receptor Er and is arguably difficult to calculate since it requires explicit solvation of the active site and more importantly it s a constant number from eq 3 the following approximation can be made AER Ep const 5 uDock 2 5 Page 12 of 18 www shadnia com 3 AEL The ligand deformation energy this can be calculated by subtracting the energy of ligand in complex by that of the free ligand according to eq 3 Two more parameters are required to describe the binding energies 4 AGgoly The ligand de solvation energy This can be calculated using Gaussian etc 5 TAS Currently there is no easy way to calculate this When all ligands in a dataset cause extraction of same number of water molecules from the active site this term can be relatively constant In other c
Download Pdf Manuals
Related Search