MTtbox documentation: Difference between revisions

From BanghamLab
Jump to navigation Jump to search
No edit summary
No edit summary
 
(9 intermediate revisions by the same user not shown)
Line 1: Line 1:
[[Software#MTtbox|Return to Bangham Lab Software]]<br><br>
[[Software#MTtbox|Return to Bangham Lab Software]]<br><br>
=Why?=
=Why?=
The aim is to model the growth of microtubules (and other dynamic organelles such as actin). Organelles grow through chemical reactions and growing organelles can collide with other organelles and membranes. To address these two features we adopted a data structure that stores microtubules as a list of vertices each of which is associated with a volume. Volumes are represented by a three dimensional array (lattice) of voxels. Regions within this volume are designated by numerical labels, e.g. 0 for cytoplasm, -4 for plasma-membrane. The size of the entire volume determines the resolution of both the collision detection system and chemical reaction/diffusion system. Resolution increases with the number of voxels. Increasing the number of voxels decreases the speed of computation and increases the demand for memory (>=16 Gbytes memory is highly desirable).  
The aim is to model the growth of microtubules (and other dynamic organelles such as actin). Organelles grow through chemical reactions and growing organelles can collide with other organelles and membranes. <br><br>
=How?=
To address these two features we adopted a data structure that stores microtubules as a list of vertices each of which is associated with a geometrically specified volume. Cell volumes are represented by a three dimensional array (lattice) of voxels. Regions within this volume are designated by numerical labels, e.g. 0 for cytoplasm, -4 for plasma-membrane. The size of the entire volume determines the resolution of the chemical reaction/diffusion system. Resolution increases with the number of voxels. Increasing the number of voxels decreases the speed of computation and increases the demand for memory (>=16 Gbytes memory is highly desirable). Dynamic organelles, such as microtubules, are represented as geometrical objects: tubes with hemispheric ends. These can collide with other microtubules, organelles and membranes. The most CPU time consuming step is collision detection.
==Current Status==
==Current Status==
MTtbox is currently under test and further development<br>
MTtbox is currently under test and further development<br>
The main data structure is called: 'data'. It can be accessed from the Matlab command line by declaring data to be global.
[http://rico-coen.jic.ac.uk/LabGuide/index.php/Modelling_using_MTtbox Pre-release internal documentation]
global data
at any time. The following documentation will refer to fields in data. It also refers to the custom menu items by ''menu:name''.
<br><br>
The MTtbox graphical user interface (GUI) was created using the rapid prototyping system:'' DArT_Toolshed\ToolBag\Demo of JRK GUI\GuiDemo.m''. This uses a text file ''GuiDemoLayout.txt'' to specify the GUI.  The GUI has a control panel (handle: ''data.PanelH'') and a graphics panel (handle: ''data.plotprops.AxesH'') The MTtbox control panel is specified by  [[MTtboxLayout.txt|''MTtboxLayout.txt'']].
=First view of the MTtbox=
===1 A===
{| border="0" cellpadding="5" cellspacing="3"
|- valign="top"
|width="300pt"|The toolbox is launched with the command
MTtbox
Which will cause the interface to appear at the top left of the monitor.<br>
The left panel (accessed through the handle: ''data.PanelH'') provides control and the right panel (''data.plotprops.AxesH'') displays the output. They can be dragged anywhere and returned to the top-left using ''menu:View:Controls to top-left''.<br><br>
 
|width="700pt"|[[Image:MTtboxUserInterface.png|400px|MTtbox GUI]]
|}
===1 B===
{| border="0" cellpadding="5" cellspacing="3"
|- valign="top"
|width="500pt"|A '''default project''' is created by selecting: ''menu:File:New Project''<br><br>It forms a cell bounded by regions labelled: Outside, cell_wall, plasma_membrane, cytoplasm and vacuole. These are concentric volumes. Length is measured in microns (time in seconds). The axis labels indicate the thickness of each layer in terms of voxels in this particular model. This is extremely low resolution - it means that the microtubules will have to be unrealistically large. This is convenient for debugging and illustrating the system but not for simulations. The outer surface of each region is coloured, e.g. (Fig. on right) the vacuole is yellow and the cytoplasm is pale green.<br><br>
The cell can be rotated etc. using the panels at the top of the display panel. Fig. below: all the regions have been hidden (uncheck each item in menu:View) and the mesh associated with the cytoplasm outer surface is displayed (check ''menu:View:Organelle meshes'')<br>
[[Image:MTtboxCytoplasmMesh.png|200px|MTtbox GUI]]
|width="400pt"|[[Image:MTtboxDefaultProject.png|300px|MTtbox GUI]]
|}
===1 C===
{| border="0" cellpadding="5" cellspacing="3"
|- valign="top"
|width="500pt"|A '''project is saved''' by selecting: ''menu:File:Save as''<br><br>Having first saved a project a default '''Interaction Function is created by selecting ''Edit'''''. A default project file contains lots of comments to provide help on how to develop the project.<br><br>At present the Interaction Functions is not copied to the new project on each Save as command - this has to be done manually.
|width="400pt"|(Comments are in green - this web version of the matlab file is created using webify_interaction_function('MT_Edinb_20120427.m')).<br><br>
The default file is largely a copy of''' MTtbox_BoilerPlate.txt''' which should be updated to reflect the latest ideas on how to build the function.
<br><br>
[[Default MTtbox interaction function|Initial MTtbox interaction function is shown here]]
|}
 
----
 
[[Image:SomeMTtboxCellShapes.png|700px|MTtbox GUI]]<br>
A) Coarse resolution (15k voxels) cubic cell. B) Coarse resolution spherical cell. C) Coarse resolution brick shape. D) Medium resolution (1M voxels) sheet of cytoplasm with plasma-membrane on one side. Blue spots are new microtubules nucleated in the plane of the cytoplasm. (Black bar: 1 micron scale bar.) Yellow: vacuole, green: cytoplasm, pink: plasma-membrane.
=Modifying a default model using the Graphical User Interface=
===2 A Changing organelles in the cell===
{| border="0" cellpadding="5" cellspacing="3"
|- valign="top"
|width="500pt"| ''menu:Organelles'' shows a list of organelles, check those that are required and then re-establish the working volumes (used for collision detection) by using ''menu:Prefs:Cell size and shape''<br>
|}
===2 B Changing working volumes used for collision detection and representing the concentrations of factors in each region===
[[Current data structures|Current data structures]]<br><br>
{| border="0" cellpadding="5" cellspacing="3"
|- valign="top"
|width="500pt"|  ''menu:Prefs:Cell size and shape'' establishes the shape of the cell and the arrangement of static organelles. The data structure it creates underpins the collision detection system.
data.cellprops.Vol
is a volume filled with labels (range 0 to -7) representing regions: not-cell, cell-wall, plasma-membrane, cytoplasm, vacuole, etc. It is re-formed whenever the cell is redefined with pushbutton ''Initialise''.<br><br>
data.working.Vol
is a copy of data.cellprops.Vol which also contains regions representing dynamic organelles (microtubules and actin). It is re-zeroed by ''Restart''. There are further volumes that supplement these two that are used for collision detection. Individual microtubule regions are recorded in
data.cellprops.microtubules.Vol
each microtubule region is represented by a unique ID.<br><br>
Factors and their associated diffusion constants are represented by
data.factorprops.Concentration
data.factorprops.DiffusionConst
each column of which represents a factor - they must be reshaped to the same format as the volume data.
 
|}

Latest revision as of 13:11, 11 December 2012

Return to Bangham Lab Software

Why?

The aim is to model the growth of microtubules (and other dynamic organelles such as actin). Organelles grow through chemical reactions and growing organelles can collide with other organelles and membranes.

How?

To address these two features we adopted a data structure that stores microtubules as a list of vertices each of which is associated with a geometrically specified volume. Cell volumes are represented by a three dimensional array (lattice) of voxels. Regions within this volume are designated by numerical labels, e.g. 0 for cytoplasm, -4 for plasma-membrane. The size of the entire volume determines the resolution of the chemical reaction/diffusion system. Resolution increases with the number of voxels. Increasing the number of voxels decreases the speed of computation and increases the demand for memory (>=16 Gbytes memory is highly desirable). Dynamic organelles, such as microtubules, are represented as geometrical objects: tubes with hemispheric ends. These can collide with other microtubules, organelles and membranes. The most CPU time consuming step is collision detection.

Current Status

MTtbox is currently under test and further development
Pre-release internal documentation

Retrieved from "http://cmpdartsvr3-v.uea.ac.uk/wiki/BanghamLab/index.php?title=MTtbox_documentation&oldid=6100"