Curved cilia give wrong cilia length

[lbozal]

Hi!! I was just measuring some cilia in cells and I realized that length measurements make sense when cilia are straight but not when they are curved. Is this a common issue? I see that the skeleton of the cilia are correct.

[hansenjn]

Hi,
Thank you for bringing this up.
No I haven't heard about this before. Could you explain a little bit more what your image properties look like and how you access the accuracy of the curved cilia measurements?
Are your images calibrated with equal x and y pixel-lengths?
If you could share an example image this would also be very helpful to troubleshoot the issue.

[lbozal]

Hi!! Find attached the original picture, the skeleton and the text files containing info and results!
let me know if everything was OK! Thanks for all the support!!

original pic_CQP_ed_CQ.txt
original pic_CQP_ed_CQl.txt
original pic_CQP_ed_CQs.txt
[original pic_CQP_ed_CQ_SKL_info.txt](https://github.com/han
original pic_CQP_ed.txt
senjn/CiliaQ/files/11212956/original.pic_CQP_ed_CQ_SKL_info.txt)

[lbozal]

Here it is the original picture!

[lbozal]

And here the skeleton

[hansenjn]

Dear @lbozal,

sorry for my very late reply. I got too busy with some other projects and lacked the time to look into the problem.

I have now released an improved version of CiliaQ that gives better length estimates than the previous versions, especially in 2D images. Maybe you could try whether this version gives you a better result and let me know, whether that helps.

Best wishes

[hansenjn]

Testing the new version with your image, the results seemed accurate in my hands.

[lbozal]

Good afternoon!! - We measured 3 different conditions: WT, MIB1KO, USP21KO - We did the CiliaQ analysis, which provided the SKL.tif files and the .txt files with the data (find this material in the link) - We represented the “cilia length [inch] (largest shortest path of largest skeleton)” parameter given by CiliaQ. The graphs by Prism had some empty spaces and did not make sense to us, as we expected a distribution of lengths along the Y axis. - We also took manual measurements with the segmented line tool in FIJI and represented them and we finally fed the SKL images to the ALPACA program. Both manual and ALPACA gave similar results, which were not similar to CiliaQ. Find the comparison of the 3 graphs in the link. Link: https://we.tl/t-qptki2ElQ0 From: Jan Hansen ***@***.***> Sent: 07 May 2023 13:28 To: hansenjn/CiliaQ ***@***.***> Cc: Laura Bozal Basterra ***@***.***>; Mention ***@***.***> Subject: Re: [hansenjn/CiliaQ] Curved cilia give wrong cilia length (Issue #13) CAUTION: This email originated from outside the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe. Dear @lbozal<https://github.com/lbozal>, sorry for my very late reply. I got too busy with some other projects and lacked the time to look into the problem. I do not clearly see the issue you mentioned. There may be tiny differences in length estimates when manually measuring the skeleton lengths to what CiliaQ measures, since CiliaQ upscales the image and blurs it for improved detection of the skeleton. Could you point me to one of these example cilia, where you think that the length is inaccurately measured and to another one where you think the length is accurately measured? And could you draw/explain exactly how you measure the different cilium length compared to how CiliaQ measures it? Do you measure the actual length in the original image or in the skeleton? Besides that I have now also released an improved version of CiliaQ that gives even better length estimates than the previous versions, especially in 2D images. Maybe you could also try whether this version gives you a better result. Best wishes — Reply to this email directly, view it on GitHub<#13 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ATL5S34JW4GGNXRBJCVA453XE6BLVANCNFSM6AAAAAAW2RTJGI>. You are receiving this because you were mentioned.Message ID: ***@***.******@***.***>> [ASOCIACION CENTRO DE INVESTIGACION COOPERATIVA EN BIOCIENCIAS] [Center of Excellence Severo Ochoa] Center of Excellence Severo Ochoa (2023-2027) All information and attachments included in this email are confidential and intended for the original recipient only. You must not share any part of this message with any third party. If you have received this message by mistake, please reply to let us know immediately, so we can make sure such a mistake does not happen again and delete this message from your system. We have put our efforts into ensuring that this message is free of errors and viruses but, despite our efforts, you should always scan all emails for any threats with proper software, as the sender does not accept liability for any damage inflicted by viewing the content of this email. We also would like to inform that all personal data included in this email or in the attached documentation, could be processed by ASOCIACION CENTRO DE INVESTIGACION COOPERATIVA EN BIOCIENCIAS with the aim of managing the legal relationship between the parties. If you would like to access, modify or oppose to the treatment of your data, you can contact us in writing. Please consider the environment before printing this email

[hansenjn]

Hi @lbozal,

thank you so much for your help, for reporting and helping to find the issue. And also for being so cooperative in sharing your data. That helped a lot. I now have found the problem. Sorry that I did not clearly see it yesterday, now I see it.

The old CiliaQ versions (v0.1.4 and earlier) could apparently not accurately enough measure the length in your type of image (2D, low pixel length of 0.003 µm/px). Due to the way the length measurements were implemented in the previous versions, the measurements had a high variance for 2D images, especially with low pixel sizes. I tried to explain it in more detail in the release notes for the newest release: https://github.com/hansenjn/CiliaQ/releases/tag/v0.1.7. It is unfortunately very complicated since this was relating to the way the length measurements were implemented programmatically. Since I have not had images with that conditions before and mostly worked with 3D images for developing CiliaQ, I had never seen the problem before but I understand it now. So it was very valuable to have your image and test the new version with this.

I have worked on improving the length measurements and now released the updated version that has no longer the issue and gives more accurate length measurements. If you install v0.1.7 you should no longer have the problem and get accurate results.

Thank you very much for your help in making CiliaQ better! Don't mind reaching out again if you encounter more issues! Getting user feedback is very valuable to find and fix issues.

[lbozal]

Thanks a lot!! I will definitely try it and come back to you!! From: Jan Hansen ***@***.***> Sent: 08 May 2023 15:00 To: hansenjn/CiliaQ ***@***.***> Cc: Laura Bozal Basterra ***@***.***>; Mention ***@***.***> Subject: Re: [hansenjn/CiliaQ] Curved cilia give wrong cilia length (Issue #13) CAUTION: This email originated from outside the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe. Hi @lbozal<https://github.com/lbozal>, thank you so much for your help, for reporting and helping to find the issue. And also for being so cooperative in sharing your data. That helped a lot. I now have found the problem. Sorry that I did not clearly see it yesterday, now I see it. The old CiliaQ versions (v0.1.4 and earlier) could apparently not accurately enough measure the length in your type of image (2D, low pixel length of 0.003 µm/px). Due to the way the length measurements were implemented in the previous versions, the measurements had a high variance for 2D images, especially with low pixel sizes. I tried to explain it in more detail in the release notes for the newest release: https://github.com/hansenjn/CiliaQ/releases/tag/v0.1.7. It is unfortunately very complicated since this was relating to the way the length measurements were implemented programmatically. Since I have not had images with that conditions before and mostly worked with 3D images for developing CiliaQ, I had never seen the problem before but I understand it now. So it was very valuable to have your image and test the new version with this. I have worked on improving the length measurements and now released the updated version that has no longer the issue and gives more accurate length measurements. If you install v0.1.7 you should no longer have the problem and get accurate results. Thank you very much for your help in making CiliaQ better! Don't mind reaching out again if you encounter more issues! — Reply to this email directly, view it on GitHub<#13 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ATL5S32GEMY574OHDUAWNODXFDU57ANCNFSM6AAAAAAW2RTJGI>. You are receiving this because you were mentioned.Message ID: ***@***.******@***.***>> [ASOCIACION CENTRO DE INVESTIGACION COOPERATIVA EN BIOCIENCIAS] [Center of Excellence Severo Ochoa] Center of Excellence Severo Ochoa (2023-2027) All information and attachments included in this email are confidential and intended for the original recipient only. You must not share any part of this message with any third party. If you have received this message by mistake, please reply to let us know immediately, so we can make sure such a mistake does not happen again and delete this message from your system. We have put our efforts into ensuring that this message is free of errors and viruses but, despite our efforts, you should always scan all emails for any threats with proper software, as the sender does not accept liability for any damage inflicted by viewing the content of this email. We also would like to inform that all personal data included in this email or in the attached documentation, could be processed by ASOCIACION CENTRO DE INVESTIGACION COOPERATIVA EN BIOCIENCIAS with the aim of managing the legal relationship between the parties. If you would like to access, modify or oppose to the treatment of your data, you can contact us in writing. Please consider the environment before printing this email