Motility module: Tracking motility of frog sperm cells not debris

[Alex Price]

Hello everyone,

I’m Alex Price, a PhD student at the University of Houston, US. I am working with frog sperm for my dissertation, in this case Xenopus leavis.

I am attempting to use the attached parameters (in the linked google drive folder below) to quantify sperm motility, but it appears that the program is tracking these bright orbs (presumably debris of some kind) as opposed to sperm cells. I obtained minimum and maximum sperm area by measuring sperm area after thresholding the first frame of my video. I haven’t put too much thought into the other parameters yet, as I am early on in the process and I am simply trying to find a good way to track my sperm.

I wonder if it has to do with Xenopus having very elongated sperm cells. I am using Xenopus testes macerates filtered through a 70 micron filter to remove debris, but I figure these debris can pass through the filter because they’re about the same size as sperm.

How can I get OpenCASA to track the motility of the sperm cells as opposed to debris? Is there a way to use ImageJ to ignore these bright debris while keeping the sperm cells (perhaps the remove outliers function?). I’ve messed around a bit with some of the functions under the “Process” tab, but I haven’t had any luck so far.

I’ve attached my input video and the output video in the folder. I am using positive phase contrast microscopy with a 10x objective (Leica DM750) and a 60 fps camera with 1.55 x 1.55 μm sensor size (Leica Flexcam i5). I am using ImageJ 1.54g and OpenCASA v2.0 on a laptop with windows 11 and 8.00 GB of RAM

Thank you all for your time and this great tool!

Sincerely,

Alex

https://drive.google.com/drive/folders/10VXXY2Br51H3k0TGqk_HGBcq87lBWo9S?usp=drive_link

[Felipe Martínez-Pastor]

Hi Alex,

Very interesting sample. I do not have the opportunity to play with this kind of samples very often.

I believe the main problem is that the scale is not set correctly. You have entered 0.155 instead of 1.55.

Then, you should adapt the other parameters. I suggest area 5-25, min VCL by 5 and max. displ. around 14 (this is important).

I see that the problem is rather the head centroid not being defined correctly, some trajectories are duplicated, maybe due to the refringency and shape of the sperm heads, I wonder if anybody has a solution for that. The round cells are not problematic once you adjust the area.

I believe that OpenCASA was designed for species with round/eliptical heads and mostly homogeneous brigthness. The analysis of other species should be possibly adjusting the parameters, but it should benefit from specific programming regarding head detection. My results were pretty acceptable after applying those numbers, though.

Merry Christmas!

Felipe

[Alex Price]

Hi Felipe,

Thank you so much for the response! I hope you had a great Christmas.

I tried the motility module again with the parameters you mentioned (see the original google docs folder for second attempt parameters), but I still had no luck as the software only tracked a few debris (output video also in the folder) which is pretty difficult to see without zooming. Perhaps its an issue with my imagej version (1.54g)? In case I missed something, could you share a screenshot of the motility parameters that you had success with for my video?

You also mentioned that my scale (.155 microns per pixel) was incorrect. However, it is my understanding that to obtain microns per pixel one must divide the sensor size by the objective zoom. In my case I I have 1.55 x 1.55 μm camera senor size and 10x zoom (1.55 / 10 = .155). Could you elaborate on why 1.55 μm per pixel is correct? I must be missing something. Would it also be wise to empirically determine my scale using a stage micrometer?

Thank you again for all your help!

Sincerely,

Alex