How to increase the accuracy of cell length measurement

Dear all,

I have recently found the measurement of cell longitudinal length is not accurate. As you can see the marked cell in my heatmap appeared 40 um but absolutely, it is much shorter than it should be.

Do you have ever met this kind of condition and how to fix it up? Thanks in advance.

snapshot.bmp (4.1 MB)

How are you estimating the length of that cell? That sample looks pretty bent, and that cell is pretty thin. You might be able to do it with the PCA tools.

Dear Richard,

I estimated it by the heat map scale and I have checked it from different view. I guess its length should not be such different with others. And what is the PCA tools? Can it also used to measure the cell length and width?

Best regards

atrichoblast cell width.bmp (4.1 MB)

If you use that line for those cells, then not many are going to be very accurate, they are too thin and bendy.

The PCA can find the long axis of the cell, it is in Mesh/Cell Axis/Shape Analysis

Dear Richard

I have tried, it worked. But I am still confused about it. At what kind of condition should I use line point to adjust the cell mesh and when I should use the shape axis function.

Best regards.

I have loaded the cell mesh and run mesh/cell axis 3D/shape analysis/compute shape analysis 3D and then run display shape axis 3D. The mesh showed below. I wondered if I want to measure the cell length and cell width, how should I do the next step. As I have tried heatmap/shape3D, but nothing happened.1-0000.bmp (4.1 MB)

In Mesh/Cell Axis/Shape Analysis/Display Shape Analysis you can select what to display for the Heatmap, so in this case you probably want Max.

Dear Richard

I have tried to run “Display shape analysis” first, but it showed “no cell axis stored in the mesh”, what should I do? I put the mesh google drive linker below, could you please help me check it? Thanks in advance.

Best regards.

Hi lei li,
You should be able to run the “Display” process after you have ran the “Compute” one.

The reason why the “traditional” cell length gives you weird values is, that it measures the cells by using a straight line originating from the cell’s centroid. If your cell is long, thin and bent it will fail. But as Richard said, the PCA (Shape Analysis 3D) should be suitable for the problem.


Dear Soeren,

Indeed the PCA can provide a more accurate result, but I suppose the value does not represent the real length. What can I do to get real length data? Thanks in advance.

Best regards.

Lei Li