Vision Based Scalable Workflow Development using Web Services

1. Vision Based Scalable Workflow Development using Web Services Jane Ren

2. Overview • Introduction to Using Web Services with Vision • ADT Library Nodes • Virtual Screening Workflow • Things to Know about Vision • Hands On Exercises • Using PrepareReceptor from Adt lib • Running virtual screening workflow • Using VirtualScreening from Adt lib • Building PDB2PQR macro / library node • Building Autodock workflow • Builidng NAMD workflow

3. Mac Users… • Make sure your Keyboard & Mouse settings are configure this way, so that your mouse can act like a Windows mouse. • Apple -> System Preferences… -> Keyboard & Mouse

4. Loading Web Services Library

5. Loading ws.nbcr.net Web Services After loading the ws library, the new nodes for Web Services appear here. These are some of the web services from http://ws.nbcr.net

6. ws.nbcr.net Web Services • The web services you just loaded are exactly the same as the ones on http://ws.nbcr.net. • http://ws.nbcr.net/opal2/GetServicesList.do

7. Loading the ADT Library

8. ADT Library Nodes for Virtual Screening New Adt lib nodes appeared

9. ADT Lib Node documentation • You can put your cursor over a node to see its documentation.

10. PrepareReceptor Node • The PrepareReceptor node creates the PDBQT file based on the input PDB file. Input - path to PDB file Output - PDBQT file (can be URL or local path)

11. Inside PrepareReceptor Node • Two Major Steps • Run PDB2PQR web service. • This step is skipped if an appropriate PQR file exists on the local machine. • Run PrepareReceptor web service. • Output is URL to PDBQT • PDB2PQR and PrepareReceptor are skipped if an appropriate PDBQT file exists on the local machine. • Output is PDBQT path on local machine. Macro that runs PDB2QR web service. Macro that runs PrepareReceptor web service

12. PrepareReceptor Web Service • Open the PrepareReceptor parameter panel • Click on PrepareReceptor • Click on PrepareReceptorWS • Double click on PrepareReceptorOpalService_ws_nbcr_net • Parameter C (preserve all input charges. ie do not add new charges, default is gasteiger charges) is NOT checked by default with MGL Tools release 1.5.6 revision 32. This may change in the next MGLTools release. • The previous PDB2PQR step is used to add hydrogen charges. • You should check/Uncheck this option based on your needs.

13. ComputeGrids Node • ComputeGrids prepare sthe GPF with the ligand library and then runs the Autogrid web service. • https://www.nbcr.net/pub/wiki/index.php?title=Virtual_Screening_Libraries contains information about the current ligand libraries on kryptonite.nbcr.net. ligand library name Receptor PDBQT file GPF template URL to Autogrid results

14. Inside ComputeGrids Node • Check if GPF template path is valid. • Exit if not • Run PrepareGPF web service based on the template GPF. • Run Autogrid web service

15. AutodockVS Node • Run Autodock virtual screening with the ligand library. Tarball of Autogrid results. This can be a URL also. Ligand library name DPF template URL to Autodock virtual screening results

16. Inside AutodockVS Node • Check if DPF template path is valid. • If not, then exit • Run Autodock virtual screening web service.

17. Virtual Screening Node • This VirtualScreening node runs the entire virtual screening experiment. It is simply a combination of PrepareReceptor, ComputeGrids, and AutodockVS. Ligand library name Input - path to PDB file GPF template DPF template URL to Autodock virtual screening results

18. Inside the VirtualScreening Node

19. Using Vision • Types • By default, a generic node/macro node port has the type None. • It is good practice to change None to a real type such as int or string. • Right click on the port and choose edit. • Select a replacement for None from the data type drop down menu. • Click on Apply and then OK.

20. Using Vision • Immediate Run Toggle • By default, the toggle is turned on. This means you have immediate run behavior. • If you change the input of a node. Then the node may start running automatically before you click on the run button. • You can turn the toggle off by clicking on the toggle button. • You can also save your network with toggle off • Next time your play with your network it will have the immediate run behavior turned off.

21. Using Vision • If you don’t know where a node is, then simply type the node name in the search box. The node will be highlighted for a couple of seconds if it is found. This node will be highlighted for a couple of seconds. Type the node here

22. Using Vision • If the straight lines that are connecting nodes are too hard to see, then spline connections. • Click on Edit • Click on Spline Connections

23. Debugging Web Services • You can view the web service log in the Vision terminal. The log may contain helpful URLs. • If you do not have a terminal open, then you need to click on the Python IDEL shell icon from ADT or PMV.

24. Hands On Exercises • Using PrepareReceptor from Adt lib • Running virtual screening workflow • Using VirtualScreening from Adt lib • Building PDB2PQR macro / library node • Building Autodock workflow • Builidng NAMD workflow

25. Exercise - Simple PrepareReceptor • Get File Browser from the Input column in Standard library. • Browse a PDB file. • Get PrepareReceptor from Adt library.

26. Exercise - Run Entire Virtual Screening Workflow

27. Exercise - Play with VirtualScreening ligand library name Input - path to PDB file GPF template DPF template URL to Autodock virtual screening results

28. Exercise - Building a Simple PDB2PQR Macro • Edit --> Create Macro • Use PDB2PQR as macro name • Drag Pdb2pqrOpalService_ws_nbcr_net node to the macro.

29. Exercise - Building a Simple PDB2PQR Macro • Double click on the PDB2PQR Web service node • Enter output.pqr as the output_file • Right click on inFile and then click on Unbind Wiget. • This will create an input port for the web service node.

30. Exercise - Building a Simple PDB2PQR Macro 3. Get the result URL This will return a list of all URLs, including input PDB, output PQR, stdout.txt, etc. Get the first URL in the list Get URL of the directory that Contains all the output files.

31. Exercise - Building a Simple PDB2PQR Macro 4. Add an input file in the main network

32. Exercise - Building a Simple PDB2PQR Macro • Try the Web Browser Node macro Open the URL in default web browser

33. Saving PDB2PQR Macro Right Click

34. Saving the Macro • In MyDefaultLib • ~/.mgltools/1.5.6/Vision/UserLibs/MyDefaultLib/Macro • In ADT Lib • $mgl_install/1.5.6/MGLToolsPckgs/AutoDockTools/VisionInterface/Adt/Macro

35. Exercise: Building Autodock Workflow • Load http://mirume.nbcr.net/opal2 ws first. • Mirume uses the CSF scheduler and will run Autodock on multiple hosts. • Input is a URL to the virtual screening files right before the Autodock step. • http://kryptonite.nbcr.net/app1249000660642/

36. Building NAMD Workflow Number of processors namd command • Mirume runs NAMD on teragrid All input files Open stdout.txt in Web browser

37. GetInputFiles Macro Node Get absolute path Get all inputs from this directory Make this files a list

38. Useful Links • MGLTools Home page • http://mgltools.scripps.edu/ • MGL Rocks roll documentation • http://www.nbcr.net/software/doc/mgl-roll/rocks-5.0/ • See sections 4.9 and 4.10 for virtual screening documenation • NBCR 09 Summer Institute wiki on virtual screening workflows • https://www.nbcr.net/pub/wiki/index.php?title=NBCR_Summer_Institute_2009 • See section 1.7.4