Hi,

I have a few questions about/suggestions for measurements and I did not see them in the discussion forum, so I apologize if they have been brought up before (which I think must be very likely).

The thing that we want to do is to track movement of nuclei in an embryo. That requires marking the location of the nuclei and measuring the displacement. There are several issues that I have noticed.

1. Right now I have learned how V3D can put a marker in 3D, but the marker is really huge with a finite spherical size. If I want to do quantitation, that requires a finer marker. I wonder if that is possible.

2. The closest way that I come up with to get a straight line between two markers is to use the line profile, starting with labeling a marker as the 1st position, and the create a line profile from the second point to it. Then by reading the information of the "traced neuron", I can get the "total length" in terms of voxel displacement I suppose? Since the size information has already been incorporated by the confocal microscope in the image so I can potential convert that back to actual length. Would this be the best way to handle this issue?

Thanks again for writing a useful package for the community!

Best regards,
Fai

For question 1, two solutions to adjust the marker size in 3d.

(a) go to Main Menu -> File -> Adjust Preferences -> select the "3D Viewer" Tab -> adjust the "Default marker display zoon factor" check box to a smaller value
(b) go to a 3D viewer, at the top-right corner, select the "Surf/Object", change value in the "Size" checkbox.

For question 2, the answer to the "suppose" part is yes. V3D currently does not use the voxel size info to compute the physical measurements, such as cubic microns. Instead, V3D computes everything in the voxel space. But there could be some really solutions such as
(a) write a V3D plugin to seelct any two markers and compute the physical distance, with voxel resolution values
(b) save the marker info in file, and use Matlab file IO toolbox and load the data and compute
(c) resample the image (using Main Menu -> Data/Image -> ... -> Resample) so that the voxel size becomes isotropic, and then compute.

hope these help.

-hanchuan



Originally posted by Yuk Fai Leung:

Hi Hanchuan,

Thanks! I should have read the manual better for 1.

For 2, I guess the most efficient strategy for me now is something like b, to save the marker info in a file and do the conversion in spreadsheet or stat programs. Please check my logic below and I also have a follow up question.

After I put markers on the 3D view, the object manager shows very fine x,y,z coordinates for the markers, for example in a 3D stack of 512x512x30 (1 micron thick z-section), I made 2 random markers to measure the 3D displacement, I have

1 - 324.305, 154.482, 11.0502
2 - 401.183, 169.303, 2.4563

Since xy's magnification and hence the scale is known and I have the z thickness, obviously I can convert the 3D voxel displacement to euclidean distance in micron.

However I don't seem to be able to find a way to export these numbers to txt from the object manager in the object manager in the 3D view. I could only go back to the image data -> landmark manager, but in that case the coordinates of one of the markers' are rounded up.

1 - 324.305, 154.482, 11.0502
2 - 402, 169, 2

If I save the markers by the save function in the landmark control, all coordinates are rounded up in the resulting file.

1 - 324, 154, 11
2 - 402, 169, 2

My follow up questions are then, 1) Do the finest coordinates in the 3D view make the most sense for this kind of displacement analysis even though the coordinates are extrapolation from the data. If so, then 2) how can I export these numbers efficiently without the rounding up issue?

Thanks!

Fai

I am not sure I understand the first question. but for the second, you can just write a simple plugin using the getSWC() and getMarker() calls, and save to whatever file formats you like!

best,
-hanchuan

Originally posted by Yuk Fai Leung:

Thanks... what I meant in the first question was that the markers in 3D give me X,Y,Z coordinates up to 3 decimal places, which is in one sense much finer than the ones that are shown in the manager, which are all rounded up. I just wonder if the coordinates with the decimal places are the best ones to use, given the 3D image is a 3D checker box with the voxels that have defined increments.

Fai