Programming Scalable Scientific Workflows

Instructors:

Michel Sanner, Ph.D. (TSRI) Guillaume Vareille, (TSRI) Vareille, Luca Clementi, (UCSD) Clementi, Jane Ren, (UCSD) Ren, Wilfred Li, Ph.D. (UCSD)
Summer Institute 2008, San Diego CA, August 3-7 3-

TSRI

Week Overview
Day 1: Introduction to Vision (M. Sanner, G. Vareille)
Basic interactions, basic networks, ImageViewer

Day 2: Building an application (M. Sanner, G. Vareille)

Writing nodes, User Library, User panels, moving widgets, noGUI execution

Day 3: Extending Pmv (M. Sanner, G. Vareille)

Extending Pmv using Vision, Icosahedral capsid

Day 4: Web Services (L. Clementi, J. Ren, W. Li)

Building a Virtual Screening application

Day 5: (M. Sanner, G. Vareille)

MatPlotlib, Volume, Student applications, wrap-up

Day 1 Overview
Background and Motivation Installing and starting Vision Interacting with the Vision GUI
GUI elements Mouse bindings

Building Networks
Image Viewer ...

Visual Programming -Abstraction of programming syntax and data structure Programming areas where - Setting up flexible Computational nodes are assembled workflows into networks
Libraries of nodes performing simple tasks

Vision: what is different

Vision is built from software components

Vision: what is different

Vision is a software component itself

Vision: what is different

No constraining data types or data model Fully scriptable Highly interactive

Installing Vision
http://mgltools.scripps.edu/download

Installing Vision

Commercial versus Academic

The following software components have restrictions for commercial use: Mslib: Computing molecular surfaces Stride: Secondary structure assignment UTPackages: Volume rendering and iso-contouring The installer will display a GUI to pick which packages To install.

Commercial versus Academic

Check to install all 3 packages Check to not install any of the 3 packages

Select individual packages for installation

Deselect individual packages for installation

Starting Vision
The installer places 3 icons on the desktop. To start Vision double click on the Vision Icon.

Vision: GUI elements

Menus Node Library Node Category Node

Current network name Network 0

Programming Canvas for Network 0

Node Documentation
NOTE Node documentation displays as tooltips

Vision: GUI elements

Merge network from file Into current network Create new network Save network Minimal Pause run execution Detached run

Load network from file

Print network Immediate Run mode network Stop toggle Load libraries execution

Stop detached execution

Vision: GUI elements

Show execution Time GANTT chart Step to next node Display advanced Search interface

Run new network Cut, Copy, Paste In debug mode Selected nodes

Search nodes by name in loaded libraries

Libraries

Some Vision node libraries: - Standard: default Vision node library

- Pmv: Vision interface to Pmv - Imaging: Interface to the Python Imaging Library (PIL) - Matplotlib: 2D graphing library - MolKit: working with biological molecules - Volume: working with 3D regular grids of scalar data - Ipython: Vision node for using IPython parallel computing -3D Visualization: Vision interface to DejaVu - Symserv: working with point symmetry operators - Web services: Vision nodes for Opal web services servers -Adt: Vision interface to AutoDockTools (ADT)

Loading Libraries

Using the Menu

Using the button

Static list of libraries Dynamic discovery of libraries

Exercise: locate a node by name

Task: Find the Dial node in the Standard library
Solution: 1 left click in the search box (A) 2 type dial <enter>. notice how the Dial node is temporarily highlighted in the library if you missed it, type <enter again>

dial

Alternative solution: 1 left click in the Input category of the Standard Node library 2 type the letter d notice how the category scrolls to show the Dial node

Exercise: drag and drop a node

Task: Drag and drop a Dial node on the programming canvas
1 left click on the Dial node in the Input category of the Standard library 2 drag the mouse to the canvas (without releasing the button) 3 release the button where you want to place the node

Left click on node

Drag node to canvas

Release

Notes
Input ports At the top Dial Widget: Feeds data to an input port

Output ports At the bottom

Port tooltip provides name, Data type and value for output Ports

- Sometimes ports with different data types can be connected is the None data type, i.e. any Python object is accepted such ports will morph to the incoming data type - Examples of datatypes: string, int, image, list, boolean, -Input ports can be required or optional. Valid date is necessary on required ports for a node to run

Notes
The filename Input port Becomes exposed at the Top of the node, allowing A name computed in the Network to be passed to The node

File browser Widget is Bound to the filename Input port

Widget can be unbound

- Unbound widget can be rebound - Widgets are for interactive user input

Vision Node
lena.jpg lena.jpg Input Port Name: filename Type: string Required: True SingleConnection: True

Read Image

Output Port name: Image Type: image

def doit(self, filename): import Image if filename: im = Image.open(filename) if im: self.outputData(image=im)

Exercise
Task: Drag and drop a Print node on the programming canvas
1 left click on the Output category of the Standard library and type p 2 click on the print Node and drag it to the canvas and release the button

Exercise
Task: Connect the output of Dial to the input port of Print
1 left click on the Output Port of the Dial node (A) 2 drag the cursor (without releasing mouse button). A green line is drawn. 3 move cursor close to print nodes input port. When you are close the line will snap to the port 4 release the mouse button to create the connection

Port type morphed to float

Left click on output port Drag green line To input port Green line snaps to input port Release

Exercise: modifying parameters

Task: Modify the Dial value
1 click on the handle and drag the cursor OR with the cursor over the dial type numbers

Note that: - values are printed to the shell Task: Prevent network execution on new data
1 Toggle the Toggle Immediate icon in the toolbar

Modify the Dial value and note that values no longer print to the shell

Notes: node outlines

Node outlines: - red: running - orange: tried to run but missing data - sick green: node failed

Turn off outlines using Edit-> Flash nodes when run

Exercise: Run Network

Task: Run the whole network, independently of data status
1 Click on the Run button in the tool bar

Note that dial value is printed to the shell each time you run

Task: Make minimal run, i.e. run only nodes that have new input data (and their children)
1 Click on the Minimal Run button in the tool bar

Note that dial value is printed when the dial has a new value but nothing is printed if the value of the dial is unchanged

Exercise: selecting
Task: Select the dial node
1 Left click on the canvas background 2 drag the cursor to draw a box around the Dial node 3 release the mouse button to toggle nodes between selected and deselected mode

-Left click on background to clear the selection - Ctrl-A selects all

Left click on background Drag red box around nodes to select Release to toggle Dial node from deselected to selected

Exercise: move nodes and connections

Task: move the Dial and print nodes together
1 select nodes to move 2 middle click on canvas background and drag cursor 3 release the mouse button to toggle nodes between selected and deselected mode

Note: using the middle mouse button on canvas background with no selection scrolls the canvas

Exercise: move a single node

Task: move the Dial node alone
1 select both nodes 2 middle click on the node background. This node becomes temporarily selected. 3 drag cursor, the node moves along. 4 release the mouse button. The previous selection is restored.

Exercise: copy-paste parts

Task: duplicate the dial-print network
1 select both nodes. Note the copy button in the toolbar becomes active. 2 click on the copy button. Note the paste button becomes active. 3 click on the paste button. Note the pasted nodes are now selected and can be moved

Exercise
Task: delete a connection
1 right click on the connection between the pasted Dial and print nodes 2 choose delete in the menu

Exercise: deleting
Task: delete a single node
1 right click on the selected print node 2 choose Delete in the menu

Task: cut multiple nodes and connections

1 select 2 nodes 2 click on the cut icon in the tool bar

Task: delete all nodes

1 use Ctrl-A to select everything in the network 2 right-click on a selected node and chose Delete

Exercise: loading a library

Task: load the Python Imaging library (PIL)
Using the button bar 1 click on the load library icon 2 select imagelib

Exercise: Image viewer network

Task: create a network to display an image
1 locate and instantiate a Read Imagenode 2 locate and instantiate a Show Image node Note that a new window is created. It will hold the image. 3 Connect the output of Read Image to the input of Show Image 4 click on the file browser icon I the node 5 navigate to Desktop/TutorialData/frames/ and select frame0000.png

Exercise: scale image

Task: modify the network to allow scaling the image
1 locate and instantiate the scale node 2 delete the connection between Read Image and Show Image 3 connect Read Image to input of Scale and output of Scale to Show Image 4 modify scale value

Exercise: rotate image

Task: modify the network to allow rotating the image

Note how sharp edges become jagged

Exercise: rotate image

Task: fix image edges using rotation interpolation filter
1 read documentation string of the Rotate node. Notice the mention of a filter. Since the filter port is not visible it must be bound to a widget. Since the widget for filter is not visible in the node it must be in the nodes Parameter Panel 2 right-click on the rotate node and select Parameter Panel. 4 in the parameter panel choose BICUBIC

Network Items Menus

Right click on network items (i.e. nodes, ports, widgets, connections to display their menus

Node Menu
Run this node and its children Freeze this node. Frozen nodes do not run Start the node editor Displays a window containing widgets associated with this node Copy, Cut, Delete this node Replace node with a fresh copy Displays a run port on the right side of the node and a trigger port on the left side

Port Menu

Show the data on this output port Display window for inspecting data content Start port editor

Widget Menu

Display option panel Only for Dial and Thumbwheel Start widget and port editor Show port at top of node Move widget to node or panel

Connection Menu

Delete connection Hide connection Toggle connections blocking flag. Non blocking means a child can execute before the parent has

Network Menu

Exercise: browser 1
Task: Build an image browser to look at the images located in the movieFrame directory on the desktop
0 delete content of network 1 locate and instantiate a NumberedName node 2 set the directory to Desktop/TutorialData/frames 3 set the base name to frame 4 set padding to 4 5 locate and instantiate a ThumbwheelInt node outputting an integer 6 connect the output of ThumbwheelInt to the number input port of NumberedName 7 locate and instantiate a Read Image node form the imaging library 8 unbind the file browser widget from the filename input port in Read Image 9 connect the filename output of the NumberedName node to the filename input port of the read Image node 10 add the show Image node and connect the output of Read Image to the output of Show Image

Exercise: browser1
Task: Build an image browser to look at the images located in the tutorialDat/frames directory on the desktop

frame

Use Edit -> Color node by library to get node colors

Exercise: save network

Task: Save the network
1 Click on the save button on the tool bar 2 navigate to Desktop 2 replace Network0_net.py by browser1_net.py in the browser 3 click on save

The network name changed to browser1

Exercise: iterate over images

Task: make the network iterate over the file numbers
1 delete the thumbwheel node 2 locate the range node, read its documentation and instantiate one to create a list of numbers ranging from 0 to 100 3 locate and instantiate an iterate node 4 send the list created by range into the iterate node 5 connect the oneItem output port of iterate to the number input port for the NumberedName node 6 Run the iterate node
frame

Task: Save the network as browser2_net.py

Exercise: iterate over images

Task: make the network iterate over the file numbers
1 delete the thumbwheel node 2 locate the range node, read its documentation and instantiate one to create a list of numbers ranging from 0 to 100 3 locate and instantiate an iterate node 4 send the list created by range into the iterate node 5 connect the oneItem output port of iterate to the number input port for the NumberedName node 6 Run the iterate node

The frame are not found ! Once a the network is saved, relative paths in the network will be interpreted relative to the SAVED location.
6 set the directory in NumberedFrame to TutorialDat/frames 7 run the network

Exercise: save MPEG

Task: save the sequence of images as an MPEG movie
1 locate and instantiate Record MPEG Movie 2 feed the image to the first input port 3 connect the begin and end output ports of iterate to the begin and end input ports of Record MPEG Movie 4 run the iterate node

Task: Save the network as saveMPEG_net.py

frame

Exercise: Play back movie

Task: play the MPEG movie
1 locate and instantiate Play MPEG node 2 unbind the widget from its movieFileName port 3 feed the movie name from the Record MPEG node into the Play MPEG node 4 kill the MPEG player to end the Play MPEG nodes execution

Day 2 Overview
Look inside a node and modify it User library of nodes Writing a new node User panels and moving widgets Command line execution Detached execution Macro nodes

Exercise: Node editor

Task: start the node editor on the Read Image node
1 Start Vision 2 Load the imaging library 2 locate and instantiate a Read Image node 3 Right click on the nodes background and select Edit

Notes: Node editor

Edit node name Add/delete/edit Input Ports Add/delete/edit Output ports Edit widget gridding Edit nodes function Add port Delete port Start widget editor Start port editor Start widget placement editor Start code editor

Notes: Node editor

- The node name is Read Image - 1 input port called filename - 1 output port called image - Note the input port is not visible at the top of the node because it gets its data from a widget which is placed Inside the node. -Double clicking on the node toggles showing/hiding widgets in the node

Exercise: Code editor

Task: start the code editor on the Read Image node
1 check the Edit check button in the node editor

- The arguments to the nodes function are named after the Input ports - Data is output using the self.outputData(portName=value, )

Exercise: Node editor

Task: add an output port that will provide the image size
1 click on Add Output Port

- An output port called out1 is created - The port appears on the node - Code is added to the function showing how to output data on the port

Exercise: Node editor

Task: add an output port that will provide the image size
2 modify the function to output im.size on the new output port

Exercise: Node editor

Task: add an output port that will provide the image size
3 Click Apply to set the function

OK does the same as Apply and hides the Code editor Cancel hides the code editor without applying changes

Exercise: Node editor

Task: read an image and verify that the port outputs the dimensions
1 read an image 2 use the output port tooltip to look at the size

Note that the data type is None

Exercise: Node editor

Task: rename the new port size and change its data type to list
1 Click edit port checkbutton for out1 2 change the name in the port editor 3 set the type to list 4 click OK

Notes: Node editor

The ports icon has changed reflecting the list data type The port was renamed

Our modification is unchanged The example code was modified

Exercise: Node editor

Task: fix the function after renaming the out1 port
1 replace out1 by size in the function 2 click Apply 3 use ports tooltip to verify that the port name has changed

Exercise: Node editor

Task: add the ability to output a thumbnail of the image upon request
1 - Add an input port allowing to specify whether a thumbnail is wanted or not: 1.a click on Add Input Port (Note the new argument in the code editor, and the new port on the nodes icon) 1.b click on edit port for the newly added port 1.c change the name to thumbnail and the data type to boolean in the port editor 1.d click OK

Exercise: Node editor

Task: add the ability to output a thumbnail of the image upon upon request
2 Bind a check button widget to the new input port 2.a click on edit port for the thumbnail input port 2.b select NECheckButton for the type of widget 2.c select node for where to place the widget 2.d click OK

Exercise: Node editor

Task: add the ability to output a thumbnail of the image upon upon request
3 - Name the widget 3.a Click on widget grid config in the node editor 3.b Type make Thumbnail in the name field 3.c click OK

Exercise: Node editor

Task: add the ability to output a thumbnail of the image upon upon request
4 add an output port for the thumbnail image 4.a Click on Add Output Port (Note new output port on the nodes icon) 4.b Click on edit port for the newly added port 4.c change the name to thumbnail and the data type to image in the port editor

Exercise: Node editor

Task: add the ability to output a thumbnail of the image upon upon request
5 - Edit the function to use the thumbnail argument and click Apply 6 use show image to look at thumbnail

Exercise: Node editor

Task: save network with modified node
1 - use File -> Save 2 - save as modif_net.py

Task: reload network and verify all modifications

1 click on the load network icon in the tool bar 2 load the network called modif_net.py 3 verify that all added input and output ports are restored 4 verify the nodes function has the modifications

The loaded network as a trailing 1 in its name to make the name unique

Exercise: Node editor

Task: make the thumbnail size a parameter controlled by a thumbwheel widget
1 add input port with good name and type integer 2 bind a thumbwheel and place it in the node 3 modify the function to use the value provided by the port

Exercise: User library

Task: save the modified Read Image node in our own library
1 delete modif1_net.py and I nmodif_net.py 2 right-click on the nodes background and select save as customized node 2 navigate the file browser to the Input folder of MyDefaultLib 3 edit the file name to be MyReadImage.py 4 click on Save

Notes: User library

The library is automatically loaded

The node appears in the Input category

Exercise: Writing a node

Task: write the Read Image node starting from a node template
1 locate and instantiate the Generic node 2 use the node editor to re-create the Read Image node - add an input port called filename of type string - bind an NEEntryWithFileBrowser widget and place it in the node - add the name filename to the widget - add an output port called image of type image - complete the nodes function 3 save node in MyDefaultLib - Input

The original Read Image node uses Image.open Image is imported in the file defining the node In your node you will have to import Image

Exercise: Looking at the source code

Task: study a source code of the node we saved in the user library
1 navigate to C:\Documents and Settings\rctraining\.mgltools\1.5.6\Vision\UserLibs\MyDefaultLib\Input 2 right-click on MyReadImage.py and select open with IDLE

Documentation string

Exercise: User Panel

Task: create a panel that provides selected widgets from a network
1 load the saved network saveMPEG_net.py 2 create a user Panel using Edit -> Create user panel (name the panel FramesToMovie) 3 use the widget menu to move the following widgets to the panel directory from NumberedName node MPEG file name from the Record MPEG movie node 4 middle-click on widget in panel and drag to move it around

Notes: User Panel

The panels allows: - running and stopping the execution - showing/hiding the network Task: save network as saveMPEGPanel_net.py to Desktop

Exercise: command line execution

Task: run the network from the command line using Vision
1 save the network with the panel as saveMPEGPanel_net.py 2 start a DOS command 3 In the DOS window type: vision ..\saveMPEGPanel_net.py

- Use the Tab key for automatic completion - Note that the network comes up with the Vision GUI

Exercise: Network Execution

Task: run a network from the command line as a program
1 In the DOS window type: ..\saveMPEGPanel_net.py - - help or h

The help message display info about command line options including parameters that can be set From the command line.
2 In the Dos window type: ..\saveMPEGPanel_net.py

Only the parameter panels comes up

Exercise: endless loop

Task: build a network with an endless loop that will print the value of a dial if the value is positive
1 create a new network using the button bar 2 turn off immediate mode using the button bar 3 locate and instantiate a while node and set the condition to 1 4 locate and instantiate a pass node 5 connect the output of while to the pass node 6 right click on the pass node and select show special ports 7 locate and instantiate a Dial node 8 right click on the Dial node and select show special ports 9 connect the trigger special port of Pass to the run special port of Dial 8 locate and instantiate an If node and set the condition to value > 0.0 10 connect the dial output to the value in put port of the If node 11 locate and instantiate a print node 12 connect the _if output port of the If node to the print node 13 turn immediate mode back on 14 save network as endlessPrint_net.py

Exercise: endless loop

Task: run the network in Vision
1 click on the run network button in the tool bar 2 make the dial positive to print to the shell or negative to stop printing

Exercise: endless loop

Special ports: - Left side of the node (run) receives signal - Right side (trigger) sends a signal after running Allow to trigger Dial execution each Time Pass runs without actually Passing data from Pass to Dial Task: run the while node When the dial is < 0.0 the print node does not flash, when you move the dial to a positive value it flashes Task: stop network execution

Notes: detached execution

Exercise: detached execution

Task: run the network in a separate process
1 click on the run detached button in the tool bar 2 make the dial positive to print to the shell or negative to stop printing

Task: stop separate process

Macro nodes
- Macro nodes represent a network as a single node in a parent network - Data can be passed into the macro and come out of it - Macros can be nested - Macros can be added to libraries of Vision nodes

Task: load the MolKit library

Macro nodes
Task: use a Lines Macro to display a molecule
1 locate and instantiate a Read Molecule node 2 locate and instantiate a Lines Macro node

The macro output date of type geometry defined in the 3D Visualization library which is pulled in automatically
3 locate and instantiate a Viewer node

A window associate to the Viewer node (black window) is created. It can be used to display 3D geometry.
4 connect the Read molecule output to the macro input 5 connect the macro output to the Viewer 6 read the molecule TutorialData/2plv.pdb from the desktop 7 click in the 3D viewer (black window) and type the letters r n c d 8 use middle mouse button to rotate the molecule

Macro nodes
Task: open the macro node
1 double click on the macro node

A new network called Lines Macro is available.

2 double-click on all expanded nodes to collapse them (hide widgets)

Macro networks have 2 special nodes input Ports and output Ports. These nodes allow data to enter and exit the macro.

Task: close the macro node

1 double click on the input Port node

Macro nodes
Task: create a new macro node
1 Edit -> create macro or Ctrl-m 2 name the macro MyMacro

You are automatically taken inside the macro network

3 locate and instantiate a pass node inside the macro 4 connect the first output port of the Macro input Port node to the input of pass 5 connect the output of pass to the first input port of the Macro output Port node Parent network

macro network

Macro nodes
Task: create a new macro node
In the parent network: 1 instantiate a Dial node and send the value into the macro 2 instantiate a print node and connect the output of the macro

Day 3 Overview
Introduction to Pmv Building a viral capsids Running PMV commands in a Vision network Creating new PMV commands

Starting Pmv
The installer placed 3 icons on the desktop. To start Vision double click on the Vision Icon.

Exercise: start Pmv

Task: load a protein into PMV
1 start Pmv 2 load the molecule hsg1.pdbqs is located in Desktop/TutorialData File -> Read Molecule

Exercise: Pmv mouse

Task: learn PMV mouse bindings
1 see page 10 of the PMV tutorial and read section 3. Mouse Bindings

Button Mod
None Left Middle Right Translate left/right (X) and up/down (Y)

Pick

Rotate

Shift

Select

Scale or Zoom

Translate in/out (Z)

Exercise: bindingsPmv key bindings

Task: learn PMV viewer keystrokes
Key
R Action

Reset view Normalize scale so all visible molecules fit in the Viewer Center on the center of gravity of all the molecules Toggle on/off Depth-cueing (blends molecule into background farther away) Toggle between transform root (i.e. scene) and transform the Viewers current object Toggle on/off Auto Depth-cueing (turn on/off setting bounding box of fog to the current object) Toggle on/off Opengl Lighting (turns on/off photorealistic lighting)

Exercise: Pmv dashboard

Task: learn undock/dock dashboard
1 click on the dock/undock dashboard button in Pmv/s toolbar

- The dashboard detached and appears in its own window

- The toolbar button is depressed while the dashboard is undocked - To re-dock the dashboard click on the button again

Notes: dashboard

Notes: dashboard
show/hide molecule display display display spheres sticks & balls lines display select ribbon display molecular surface

representation commands

color commands

Notes: dashboard
Atom: by atom type RAS: Rasmol amino acid colors SHA: Shapely amino acid colors SecStr: by secondary structure Inst.: by instance

color commands

Exercise: Pmv dashboard

Task: Display chain A as CPK
1 left-click in dashboard on the circle at the intersection of chain A and CPK

Exercise: dashboard
Task: display chain B as scaled down CPK
1 right-click in dashboard on the circle at the intersection of chain B and CPK

The option panel of the display CPK command appears

2 Set scale Factor to 0.3 3 click OK

Exercise: dashboard
Task: un display CPK spheres for chain B un display lines for chain A and B display ribbon for chain B
1 left-click at the intersection of chain B and CPK 2 left-click at the intersection of chain hsg1 and Lines 3 left-click at the intersection of chain B and Rib.

Exercise: dashboard
Task: color hsg1 by atom type
1 left-click at the intersection of hsg1 and Atom

- Coloring is applied to all representations -The ribbon is blue because the N atom info is used - The color button NOT check buttons (i.e. no state)

Exercise: dashboard
Task: color chain B by secondary structure element
1 right-click at the intersection of chain B and Sec.Str.

- A geometry chooser is displayed allowing to select the representations that will be colored

2 select secondarystructure 3 click OK

Exercise: dashboard
Task: display residue Arg8 in chain B as sticks and balls and color the sticks and balls by DG
1 expand chain B 2 left-click at ARG8 and S&B 3 right click at ARG8 and DG 4 select sticks and balls 5 click OK

Exercise: Viral capsid

Task: build the viral capsid of Polio Virus
1 right-click on hsg1 in the dashboard and select Delete 2 right-click on PMV Molecules and load Desktop/TutorialData/2plv.pdb 3 start Vision by clicking on the Vision button in the toolbar

The Pmv node library has Vision nodes specific to PMV

Exercise: Viral capsid

Task: build the viral capsid of Polio Virus
4 load the symmetry server library of Vision nodes

Exercise: Viral capsid

Task: build the viral capsid of Polio Virus

5 build the following network

6 run the network

Exercise: Viral capsid

Task: change the representation from lines to a coarse molecular surface
1 using the dashboard un-display the lines for 2plv 2 execute the command Compute -> Coarse Molecular Surface from PMV menu using default parameters 3 - place cursor on 3D viewer window and type r, n, c

The compute coarse molecular surface command is implemented as a Vision networks that is loaded the first time the command runs

Exercise: Viral capsid

Task: color the surface by chain
1 using the dashboard color 2plv by chains

Exercise: Viral capsid

Task: color the surface by depth in the capsid

Exercise: Viral capsid

Task: expand the capsid by translating each 5-fold copy along its 5-fold axis
1 double click on the Icosahedral1 node to expand the macro 2 double click in the 5-fold node to display its parameter panel 3 find the values 0.000 0.525 0.851 in the parameter panel 4 locate and instantiate a translate node 5 double click on the translate node to display its parameter panel 6 find the translation vector 1 0 0 and replace with 0.000 0.525 0.851 7 delete the connection between the 5-fold and the 3-fold nodes 8 insert the translate node between the 5-fold and 3-fold nodes 9 modify the translation length in the parameter panel of the translation node (right click on the thumbwheel to increase sensitivity to 10 for better results) 10 witness the capsid expand

Exercise: Viral capsid

Task: expand the capsid by translating each 5-fold copy along its 5-fold axis

Exercise: Viral capsid

Task: add range and iterate node to automate expansion

Exercise: Viral capsid

Task: add range and iterate node to automate expansion
1 unbind vector length widget in the parameter panel of the translate 2 add range node from 0 to 100 in steps of 5 3 add iterate node to iterate over range output 4 feed value from iteration into vector length port

Exercise: Viral capsid

Task: build the TMV capsid
1 delete the 2plv molecule 2 load the Desktop/TutorialData/tmv/2tmv.pdb protein 3 hide the lines and display a coarse molecular surface 4 instantiate Pmv Viewer node 5 instantiate a Choose Geom node 5 connect the viewer to the choose Geom (the combo box will be populated) 6 select root|2tmv in the combo box 7 instantiate a Helix node 8 Instantiate a Set Instances node (use the one from the 3D vis library) 9 connect Choose Geom and Helix outputs to Set Instances 10 double click on helix to display its parameter panel 11 set copies to 50 12 start changing the angle of the helix (values around 20 are good) 13 interact with the molecule in the viewer (r n c and rotate) 14 start changing the rise of the helix 15 set the rise to 1.43 and the angle to 22.04 16 color by Instances (using the dashboard)

Exercise: Viral capsid

Task: build the TMV capsid

Exercise: use Pmv cmds in Vision

Task: build a network to run the computeMSMS cmd
Run the MSMS command in Pmv first to see possible arguments 1 in Pmv delete 2tmv and load tf_1.pdb 2 in Pmv use the Compute -> molecular surface command note the arguments that are possible include: surface name, probe radius, density, per molecule, etc 3 click Dismiss 4 in Vision: create a new network 5 instantiate a Pmv node and select computeMSMS for the cmd: 6 instantiate a Run_command node and display its parameter panel 7 connect the cmd output port of the Pmv node to the Run_command (note what happens) 8 instantiate a tf_1 node and connect the molecule to the new input port on Run computeMSMS 9 the surface gets computed after you connect 10 vary the probe radius dial to higher values

Exercise: use Pmv cmds in Vision

Task: modify the network to compute a surface for each amino acid
1 instantiate a select MolFrag node to get a list of residues 2 send the tf_1 molecule into this node and select Residue for level 3 add an iterate node to iterate over the list of residues output by this node 4 connect the oneItem output port of iterate to the compute MSMS node

Surface is computed for the whole molecule but only the patch for the last residue is displayed

Exercise: use Pmv cmds in Vision

Task: modify the network to compute a surface for each amino acid
5 un-check the perMol check button in the computeMSMS parameter panel

Surface is now computed for the for the set of atoms in the residue

6 run the iterate node and watch the surface walk along the chain

Exercise: scale CPK radii by charge

Task: display CPK spheres scales y the atomic charge
1 delete all molecules and load indinavir 2 display CPK and color by atom types 3 create a new vision network 4 instantiate an indinavir node, Select MolFrag and Extract Atom Property 5 send the molecule into select MolFrag with level set to Atom 6 send the resulting AtomSet into Extract Atom Property and set prop. Name to charge 7 instantiate a Dial node and set it to 3.0 and an op2 node 8 in op2 select operator to be mul and check apply to elements 9 connect the list of charges to the first input port of op2 and the dial to the second 10 instantiate a Pmv Viewer node and a Choose Geom 11 connect the viewer output to the Choose Geom node and select root|indinavir|cpk 12 instantiate a Call method node and set the signature to Set radii <enter> 13 connet the output of Choose Geom to the first port of call method 14 connect the list of scaled charges coming out of mul to the second port

Exercise: scale CPK radii by charge

Task: display CPK spheres scales by the atomic charge

Exercise: new PMV command

Task: run the 3res.py script from the command line using Pmv
1 start a DOS command 2 In the DOS window type: pmv i 3res.py 1BVN_4.0_1.0_-17.0.txt

- Use the Tab key for automatic completion - Note that the network comes up with the Vision GUI

Exercise: new PMV command

Modify the parameters of ShowSolutions to browse results

Day 4 Overview
Using Web Services with Vision ADT Library Nodes
Using Existing ADT Nodes Building a ADT Lib Node

Vision Workflows
Building a NAMD web service workflow Building an Autodock web service workflow using Teragrid resources

Day 5 Overview
The Volume library The vizlib library The matplotlib library Student problems Wrap up

Working with volumetric data

Working with volumetric data

Boolean operation can be performed on masks to create complex masks Here a spherical slab mask is created by XORing 2 spheres

2D plotting
http://mgltools.scripps.edu/packages/vision/matplotlib/vision-networks-for-matplotlib

3D Visualization

3D Visualization

3D Visualization: GUI overview

Bind Mouse to transform Operations assigned to mouse buttons (changes with modifiers) When checked 3D Xforms apply to root Root geom parent of all geometries and current object Master geom for all geoms of a given molecule

Geometry objects hierarchy Reset Xform of current object Fit the scene in the view

Geoms created by Pmv cmds for that molecule Set rotation center to center of the scene

3D Visualization: GUI overview

Select property panel to show

Object property panel

3D Visualization: GUI overview

Camera property panel

3D Visualization: GUI overview

Lights property panel Clipping planes property panel

2D plotting
http://mgltools.scripps.edu/packages/vision/matplotlib/vision-networks-for-matplotlib Desktop/doc/Examples/matplotlib

Matplotlib in DejaVu

Exercise
1