ImageJ

== Introduction ==

[[wikipedia:Kymograph|Kymographs]] are a way to represent 
on a single image a dynamic process. They are particularly useful to monitor 
in a glimpse a moving organelle and characterize its motion.
They can be seen as a ''x-t'' scan, where the intensity along a given line is plotted
for all images of a stack.  Each time point gives a intensity line, plotted ''e.g.'' along the 
x axis of the kymogrpah. These lines are stacked along the y axis for all frames. 
So we get an image where we move along space in the x direction and along time
in the y direction.

Examples of usage can be found in the references listed at the end of this page.
We will introduce them in a real-world example, using the most simple functions
in ImageJ.


== Context ==

This is a movie made in the lab of Ewa Paluch, MPI-CBG, featuring a floating L929 cell 
(mouse fibroblast), stained for myosin (RLC-GFP) being ablated in the cortex.
The transmitted signal is overlayed with a confocal image of the GFP channel in various 
ways.  The acto-myosin cortex is the part of the cytoskeletton that lies just below the membrane
This cortex is active and generate a contractile force that presses on the cell volume.
At rest, this tension gives the floating cell this spherical shape.

A picosecond pulsed laser allowed us to make a micro ablation of the cortex, without
disrupting the membrane. The ablation laser power saturates the transmitted detector,
which results in white frames in the movie.

When ablated at the left spot, the constraint on volume is relaxed, and the cortex can now
contracts, expulsing some cytoplasm through the hole, like a relaxing ball that would be squeezed
by your hands.
This leads to the formation of a bleb, a woughly spherical bulge of membrane, initially devoid of
cortex. After a while, a new cortex is polymerized and the bleb retracts.
Goal of this study is to get the dynamics of the bleb growth and retraction.


== Bleb growth dynamics ==

First, get the raw transmitted movie [http://fiji.sc/tinevez/web/Bleb_growth_transmitted.tif here (22MB)].

We would like to get the speed at which the front of the bleb grows.
The simplest way to do so is to use kymographs. 
The idea is to draw a line from the center of the cell that passes through the bleb.
To asjust the line position, select a frame where the bleb is still small, and draw a line 
ROI from the center through the bleb.

[[Image:Generate_and_exploit_Kymographs-1.jpg|center]]

Check the feature using a profile plot (shortkey {{key|K}} or {{bc | Analyze | Plot profile}}).

[[Image:Generate_and_exploit_Kymographs-2.jpg|center|frame|A line profile]]

The edge feature is not that clear on this picture. At about
7 um, one can see that the dark pixels and the bright halo around the bleb front
generated some kind of ridge.

The kymograph consists in plotting this profile, encoding the intensity in 
grayscale level (or color for color images) for all slices. Rather than creating 
a plugin for this, we can use a function that was initially made for Z-stacks:
the {{bc | Image | Stacks | Reslice}} command (shortcut "/").
An option dialog appears; the defaults are fine for us, just press OK. 
We then get the kymograph:

<!--
[[Image:Generate_and_exploit_Kymographs-3.jpg|center|frame|A kymograph]]
-->

[[Image:Generate and exploit Kymographs-3bis.png|right|frame|A kymograph]]

On this image, space is encoded along the x axis, following the line ROI
from left to right. Since we drawn the ROI from the middle of the cell to
the exterior of the cell, intensities measured within the cell are reported
to the left of the image. At almost the extreme right, we get the cell's
border and the bleb front.

Time is encoded along the y direction. The first frame measurements is
reported on the top line, and the bottom line reports the last frame. 
The white band correspond to the ablation, between frame 11 and 16,
and appears between the 11 and 16 pixels of the kymograph.

ImageJ can't deal with a calibration unit that is different in x and y axes, 
which is the case here (space and time). So for further analysis,
we will have to put it right ourselves using the pixel size for the 
x calibration and the frame interval for the y calibration.
I found it safer to remove totally any calibration from the images, even 
from the source stack. This way no confusion can be arise from a delay that
would be measured in µm.


The front of the growing bleb can be seen below this white band, as 
a dark edge at the right of the image. The bleb growth appears here as 
a rather steep curve below the white band. Its retraction takes much longer,
which results in having a drak edge slowly moving towards the left part 
of the image as we screoll down the kymograph. Other features within the
cell can also be seen within its volume. In the case of this transmitted
image, they are various granules movements which we do study here. 

To get the bleb front dynamics, one can directly plot a multi-
line ROI and get its coordinates for further analysis. We could also 
try to segment it a bit automatically. 

* Use the {{bc | Process | Find edges}} command, to stress out the bleb front.
[[Image:Generate_and_exploit_Kymographs-4.jpg]]

* Threshold it.

[[Image:Generate_and_exploit_Kymographs-5.jpg]]

* Use the multi-line ROI to select points along the edge, and export their coordinates using the ''{{bc | File | Save as...}} > XY Coordinates'' command.
You now have a text files made of 2 columns, the first one for the time, 
the second one for the bleb front position, ready for analysis. Do not forget
to put the calibration right, tough. 

[[Image:Generate_and_exploit_Kymographs-6.jpg]]


Next step is to get numerical values and model for this growth.
We leave it to another step, for the image analysis part is done. 

<u>Exercise:</u> You can also think of a more clever way
to extract the front coordinates, with sub-pixel accuracy (paste your ideas in the [[Talk:Generate_and_exploit_Kymographs|discussion page]]).

== Myosin dynamics at bleb front ==

Now download [http://fiji.sc/tinevez/Bleb_growth_myosin.tif here (22 MB)] the movie for the same experiment, but measured in the fluorescence channel. It pictures myosin dynamics at the cortex.
Our goal is to measure the myosin recovery rate at the bleb front during its growth.
This would typically be complicated: we are trying to measure the intensity of something that moves over time. A sophisticated solution would be to segment the bleb front as an object and then measure the intensity given the objects. We propose here a simple solution involving a bit of manual interaction.

First,change a bit the contrast so that we can see the cortex well, so as to place correctly our line ROI. To do so, use {{bc | Image | Adjust | Brightness/Contrast}} and play with the ''Maximum'' slider. Do not press ''Apply'': we just want to improve what we see on the screen, not the actual intensities stored in the stack. Then draw a line ROI that goes - as before - from the center of the cell to the exterior, crossing the bleb grow spot.

[[Image:Generate and exploit Kymographs-7.png|center]]

Get the kymograph from this movie using the ''Reslice'' command. Thrash the calibration before doing it if you feel the use for it. You can also use the dynamic version of ''Reslice'', shipped with Fijim and that can be found in {{bc | Image | Stacks | Dynamic reslice}}. The advantage is that the kymograph is updated live when you move the ROI. After improving the contrast, it should look like this:

[[Image:Generate and exploit Kymographs-8.png|center|frame|A kymograph for fluorescence image]]

We can distinguish the bleb growth dynamics with this image, particularly on long delays when the acto-myosin cortex as repolymerized and reform a identifiable shell. It then appears at a bright band towards the bottom of the kymograph. We also have more or less our answer: along this bright band is pictured the myosin dynamics at the bleb front. To measure it accross time, we just have to plot a profile that goes from the top of the kymograph to its bottom and that follows the bleb edge.
The profile function has a nice features that allow to average over a certain line width. You can specify this width by double-clicking on the line tool button in the main Fiji window. 

[[Image:Generate and exploit Kymographs-9.png|700px|Plotting a profile on a kymograph]]


You can see that the myosin signal seems to recover very soon after ablation. But here is a gotcha that limits a bit the scope of this approach. Notice the units of the x axis of the profile. It is a distance in pixels where you expect a delay in seconds. It is anyway none of it, but an meaningless mix of both. The profile tool computes the distance that it reports in the x axis using the euclidian distance, that is 

<math>s=\sqrt{(x_2-x_1)^2+(y_2-y_1)^2}</math>

but since on our kymograph the x coordinates report the space and the y coordinates report the time, the x axis of the profile has no meaning, and can't be exploited as is.



<u>Exercise:</u> Think of a way to put this right, that is to get a profile where the x axis would be the crrect delay measured from the ablation offset (discuss it in the [[Talk:Generate_and_exploit_Kymographs|discussion page]]).

== Examples and references ==

* {{Citation | last = Dhonukshe | first = Pankaj
| last2 = Vischer | first2 = Norbert
| last3 = Gadella | first3 = Theodorus W. J., Jr
| title = Contribution of microtubule growth polarity and flux to spindle assembly and functioning in plant cells
| journal = J Cell Sci
| volume = 119
| number = 15
| pages = 3193-3205
| doi = 10.1242/jcs.03048
| year = 2006
| abstract = The spindle occupies a central position in cell division as it builds up the chromosome-separating machine. Here we analysed the dynamics of spindle formation in acentrosomal plant cells by visualizing microtubules labelled with GFP-EB1, GFP-MAP4 and GFP-alpha-tubulin and chromosomes marked by the vital dye SYTO82. During prophase, few microtubules penetrate the nuclear area, followed by nuclear envelope disintegration. During prometaphase, microtubules invading the nuclear space develop a spindle axis from few bipolar microtubule bundles, which is followed by spindle assembly. Using a novel quantitative kymograph analysis based on Fourier transformation, we measured the microtubule growth trajectories of the entire dynamic metaphase spindle. Microtubules initiating from spindle poles either pass through the metaphase plate to form interpolar microtubule bundles or grow until they reach chromosomes. We also noticed a minor fraction of microtubules growing away from the chromosomes. Microtubules grow at 10 um/minute both at the spindle equator and at the spindle poles. Photobleached marks created on metaphase and anaphase spindles revealed a poleward tubulin flux. During anaphase, the velocity of tubulin flux (2 um/minute) equals the speed of chromatid-separation. With these findings we identified spatially coordinated microtubule growth dynamics and microtubule flux-based chromosome-separation as important facets of plant spindle operation.
| url = http://jcs.biologists.org/cgi/content/abstract/119/15/3193
}}

* {{Citation | last = Zhou | first = H. Mimi
| last2 = Brust-Mascher | first2 = Ingrid
| last3 = Scholey | first3 = Jonathan M.
| title = Direct Visualization of the Movement of the Monomeric Axonal Transport Motor UNC-104 along Neuronal Processes in Living Caenorhabditis elegans
| journal = J. Neurosci.
| volume = 21
| number = 11
| pages = 3749-3755
| year = 2001
| abstract = The formation and function of axons depends on the microtubule-based transport of cellular components from their sites of synthesis in the neuronal cell body to their sites of utilization at the axon terminus. To directly visualize this axonal transport in a living organism, we constructed transgenic lines of Caenorhabditis elegans that express green fluorescent protein fused to the monomeric synaptic vesicle transport motor, UNC-104. This UNC-104:: GFP construct rescued the Unc-104 mutant phenotype and was expressed throughout the nervous system. Using time-lapse confocal fluorescence microscopy, we were able to visualize fluorescent motor proteins moving in both directions along neuronal processes, some of which were identified definitely as axons and others as dendrites. Using kymograph analysis, we followed the movement of >900 particles. Most of them moved in one direction, but not necessarily at the same velocity. Ten percent of the observed particles reversed direction of movement during the period of observation, and 10% exhibited periods of movement interspersed with pauses. During episodes of persistent movement, particles moved at an average velocity of 1.02 {micro}m/sec, which is close to the in vitro velocity of microtubule gliding driven by purified monomeric kinesin at high motor density. To our knowledge, this is the first direct visualization and analysis of the movement of specifically labeled microtubule motor proteins along axons in vivo.
| url = http://www.jneurosci.org/cgi/content/abstract/21/11/3749
| eprint = http://www.jneurosci.org/cgi/reprint/21/11/3749.pdf
}}

[[Category:Tutorials]]