Cell front migration profile from spherical body, angle/moment, area , furthest distance

Hi everyone,

I am hoping I can get some advice on the following problem. I compiled and put all my ideas/methods/approach in detail in order to verify what is the best method to analyse shape profiles (angle or orientation, amount of migration 0-180 degrees) from an internal circular body outwards.
I have spheroids migrating on different surfaces, such as standard tissue culture polystyrene, and fibres. I tried multiple methods to measure the amount and angle of cell migration. The methods I have tried gave different results. So I am here asking for clarification in terms of what is actually being measured and whether I am implementing the right methods/tools to extract the features I am looking for.
I attached a PPT showing the workflow, and please correct me if i’m wrong with the steps implemented towards quantifying the migration profile: area, and the amount/angle of migration from spherical body. The methods used include: Directionality plugin, Moment analysis, Analyze particles, Measure.
In my analysis I tried complex and simplified shape versions of the same migration profile.

Image processing

Method 1a: simple shape, Directionality [slide page # 3, 5 ]
Load image- Make 8-bit - adjust brightness- threshold- Make binary - Plugin shape analysis - Hull and circle (unclick circle)- click on hull - clear outside - invert
Method 1b: complex shape, Directionality [slide page # 4, 6]
Load image - Make 8-bit - adjust brightness - threshold - Make binary - Fill holes - Noise despeckle - select object with largest area - clear outside

Method 2: Analyze - measure
2a: complex shape [slide page # 6]
2b: complex shape [slide page # 7]

Method 3: Moment analysis A and B [slide page # 8]
3a: Complex shape
3b: Circle cleared in centre of 3a (above) to show migration from body.

Method 4 [slide page # 9 ] Amount and angle from internal circle boundary
Merge channels (Bright-field and processed imaged such as the one in Method 1b
Draw circle around bright red area in centre - clear circle - split channels
Make 8-bit

Results:
1a: direction 65.97 degrees (simple shape)
1b: direction -7.49 degrees (complex shape)
2a: Angle 167.413 degrees (complex shape)
2b: Angle 17.932 degrees (simple shape)
3a: Angle = 167.400 degrees, orientation = 12.857 degrees (complex shape)
3b: Angle = 167.400 degrees, orientation = 9.398 degrees (complex shape with cleared circle in the middle)
4a: still in progress but an attempt is shown in 3b.

Discussion:

As you can see from the results above, the angle extracted is different depending on the method. I would like some clarification on how each method is defining the measurement.
What i would like to do? I would appreciate if someone could explain and/or guide me through this
Method 4: [slide page # 9 ] Is it possible to measure from image attached as example (picture1)
Picture1

  1. Distance of cell fronts from Centroid of cleared middle circle, or circle boundary?
  2. The amount and angle/direction/orientation of cell fronts wrt. centroid of cleared middle circle, or circle boundary?
  3. The furthest distance from centroid of cleared middle circle, or circle boundary

Thank you all in advance

1 Like

I’m guessing that there is no solution/method available in FIJI to do this kind of work.

@iwanaknow There are methods available—e.g. OrientationJ (but that one is not open source). It would help to have a sample dataset upon which to base any analysis efforts. Right now we don’t even know if your data is 2D? 3D? Time-lapse? If your data is time-lapse then TrackMate might help.

1 Like

I looked into OrientationJ (havent used it though), but I found it is similar to Directionality plugin in Fiji.

@iwanaknow If you would like further advice from the community, please post a sample dataset here. You can attach ZIP files, or use a cloud sharing link such as Dropbox.

@ctrueden I just attached a zip file to this thread containing brightfield, Alexa568, and Hoechst channels.

I want to determine extent of migration from a circular position (spheroid at t=0) or from centroid coordinates (x,y), something like radial distribution in 360 degrees. Maybe its not the right term to describe it, let me know if my wording makes sense.

thanks

@iwanaknow Sorry for the delay in reply. I don’t see the attachment. Just reply and drag-and-drop the ZIP into the editor window? Or create a Dropbox or other cloud-share platform link and paste that?

Hi @ctrueden, attached brightfield, Alexa568, and Hoechst channels.

1 Like