NITRC Nutil - Neuroimaging utilities Forum: general

RE: QUINT Workflow Report Not Counting some Pixels

Sharon Christine Yates — Tue, 30 Jul 2024 7:52:44 GMT

Dear Busra,
Sorry for the late reply. The committee members concerns are valid here, there are downsides with using new analytical methods that have not been as extensively validated as more traditional methods: sources of error may differ and it can make comparison across studies more complicated. However, traditional methods such as stereology also have their limitations, with many upsides associated with use of standardised reference atlases. Semi-automated methods such as the QUINT workflow are becoming more common, we have written about this in our recent review article:
-Bjerke IE, Yates SC, Carey H, Bjaalie JG, Leergaard TB. Scaling up cell-counting efforts in neuroscience through semi-automated methods. iScience. 2023 Aug 7;26(9):107562. doi: 10.1016/j.isci.2023.107562. 
We have done our best to validate the QUINT workflow from different perspectives. This is documented in the following articles. The Gurdon et al article in particular focuses on validating registration to atlas regions. 
-Yates et al. QUINT: Workflow for Quantification and Spatial Analysis of Features in Histological Images From Rodent Brain. Front Neuroinform. 2019 Dec 3;13:75. doi: 10.3389/fninf.2019.00075. 
-Groeneboom et al. Nutil: A Pre- and Post-processing Toolbox for Histological Rodent Brain Section Images. Front Neuroinform. 2020 Aug 21;14:37. doi: 10.3389/fninf.2020.00037.
-Gurdon et al. Detecting the effect of genetic diversity on brain composition in an Alzheimer's disease mouse model. Commun Biol. 2024 May 20;7(1):605. doi: 10.1038/s42003-024-06242-1.
Good luck with your PhD.
Best regards,
Sharon Yates
 
Originally posted by busra:
<blockquote>
Dear Dr. Yates, Firstly, I would like to express my excitement in following your work closely. In my doctoral thesis, I conducted morphometric analyses on rats with Alzheimer's using the Quint workflow. However, due to concerns raised by my committee members regarding this workflow, I received feedback for revisions. I aligned images with an atlas template using QUICKNII and VisuAlign and performed volumetric analyses stereologically. Nonetheless, my committee members pointed out potential shrinkage or displacement during the histological preparation of samples, questioning the validity of manually fitting them to the template. They particularly expressed concerns about errors that could arise when placing small neuroanatomical areas onto the template. They questioned how the program distinguishes local features in areas where histological boundaries are not clearly defined, even in stained sections. They inquired about the methods used in Quint workflow to validate measurement accuracy and requested information on error coefficients or standard deviations in these measurements. Furthermore, they highlighted that tissue blocks may not perfectly align with the x, y axes during sectioning by a microtome, raising questions about how axial shifts affect the placement of atlas templates and subsequent measurements. They also asked about integrating 2D histological images into a 3D atlas and how this process ensures accuracy. Despite referencing your team’s research and providing explanations, they remained unconvinced. Could I please ask for your assistance in addressing these concerns? Your guidance in this matter would be greatly appreciated. Best regards,
</blockquote>

RE: QUINT Workflow Report Not Counting some Pixels

busra — Sun, 30 Jun 2024 14:38:15 GMT

Dear Dr. Yates, Firstly, I would like to express my excitement in following your work closely. In my doctoral thesis, I conducted morphometric analyses on rats with Alzheimer's using the Quint workflow. However, due to concerns raised by my committee members regarding this workflow, I received feedback for revisions. I aligned images with an atlas template using QUICKNII and VisuAlign and performed volumetric analyses stereologically. Nonetheless, my committee members pointed out potential shrinkage or displacement during the histological preparation of samples, questioning the validity of manually fitting them to the template. They particularly expressed concerns about errors that could arise when placing small neuroanatomical areas onto the template. They questioned how the program distinguishes local features in areas where histological boundaries are not clearly defined, even in stained sections. They inquired about the methods used in Quint workflow to validate measurement accuracy and requested information on error coefficients or standard deviations in these measurements. Furthermore, they highlighted that tissue blocks may not perfectly align with the x, y axes during sectioning by a microtome, raising questions about how axial shifts affect the placement of atlas templates and subsequent measurements. They also asked about integrating 2D histological images into a 3D atlas and how this process ensures accuracy. Despite referencing your team’s research and providing explanations, they remained unconvinced. Could I please ask for your assistance in addressing these concerns? Your guidance in this matter would be greatly appreciated. Best regards,

RE: QUINT Workflow Report Not Counting some Pixels

Sharon Christine Yates — Mon, 15 Jan 2024 13:30:37 GMT

Hi,
The "Region pixels" is the no. of pixels in your segmentations that corresponds to the region-of-interest. So if you know the pixel scale of the segmentations (area represented by each pixel in your segmentation, e.g. 10mm2 per pixel), you can enter this no. (10) and the unit (mm2) in Nutil, and the areas will be automatically calculated and reported in the "region area" field. It is usually possible to obtain the pixel scale of the original images from the microscope scanner. Assuming you downscaled the images before segmentating them, you'll have to calculate the correct pixel scale of the downscaled images before you can use the number in Nutil. 
Good luck,
Sharon
Originally posted by busra:
<blockquote>
<div class="subject">Nutil Region Area</div>
<div class="body">
Hi, Sharon, First of all, thank you and your team for this wonderful workflow. My PhD thesis is a morphometric study on rats with Alzheimer's disease and using QUINT workflow in my study was very useful. However, I would like to ask you about a few things that are on my mind. Can I accept the Reference Region Area information given in the Nutil output report as a morphometric area? My goal is to get volumetric results from this data. For this, I will use the stereology method in my study, but I wonder if this is correct.  For example, if we talk according to the table below; can I consider the cortex area as 4772968 pixels? Also, even if I convert units from pixels to mm2 during nutil quantifier, the result is still 4772968. What should I do in this case?
 
Region name    Region pixels    Region area    Area unit Cortex    4772968    4.772.967.600.046    pixels Fibretracts    1755011    1.755.010.611.384    pixels Hippo    1031737    1.031.736.514.834    pixels Olfactory    3649652    3.649.652.441.756    pixels
 
Best regards..
</div>
 
 
 
</blockquote>

RE: Customized Quantification for left vs. right hemisohere

Sharon Christine Yates — Mon, 15 Jan 2024 12:48:46 GMT

Dear Ying,
Currently it is possible to generate separate reports for each hemisphere by using hemibrain masks in Nutil (you'll have to run Nutil twice). See instructions here: https://nutil.readthedocs.io/en/latest/QuantifierMasks.html. You can quickly generate hemibrain masks corresponding to your sections by using the QuickMask or QuickMaskNL provided on the NITRC page for QuickNII or VisuAlign respectively (email me at s.c.yates@medisin.uio.no if you have problems with the tools, if needed I have a more detailed user manual I can forward you). 
In the future, we plan to incorporate hemibrain atlases in the tools. These will automatically create different reports for the left and right hemisphere, but this is still in the pipeline and will take some time. 
good luck
Sharon

Customized Quantification for left vs. right hemisohere

Ying Han — Fri, 12 Jan 2024 18:39:43 GMT

Hello,
    Thank you for the amazing tool! I have been using Quint in my Parkinson's disease project to quantify cells. I am wondering if there is a way to generate reports that also include results for each hemisphere. Thank you!
Ying

RE: QUINT Workflow Report Not Counting some Pixels

busra — Thu, 28 Dec 2023 10:10:20 GMT

<div class="subject">Nutil Region Area</div>
<div class="body">
Hi, Sharon, First of all, thank you and your team for this wonderful workflow. My PhD thesis is a morphometric study on rats with Alzheimer's disease and using QUINT workflow in my study was very useful. However, I would like to ask you about a few things that are on my mind. Can I accept the Reference Region Area information given in the Nutil output report as a morphometric area? My goal is to get volumetric results from this data. For this, I will use the stereology method in my study, but I wonder if this is correct.  For example, if we talk according to the table below; can I consider the cortex area as 4772968 pixels? Also, even if I convert units from pixels to mm2 during nutil quantifier, the result is still 4772968. What should I do in this case?
 
Region name    Region pixels    Region area    Area unit Cortex    4772968    4.772.967.600.046    pixels Fibretracts    1755011    1.755.010.611.384    pixels Hippo    1031737    1.031.736.514.834    pixels Olfactory    3649652    3.649.652.441.756    pixels
 
Best regards..
</div>
 
 
 
Originally posted by Sharon Christine Yates:
<blockquote>Hi Diana, Please send this request to EBRAINS support at support@ebrains.eu and we'll do our best to help. To investigate we'll need a copy of your .nut file, a segmentation, an atlas map and the xml of json file. One segmentation file with the corresponding atlas map should do (include the section where the problem occurs). I assume you haven't set a minimum object size (under the advanced settings) as this could explain the exclusion of small objects. Best wishes, Sharon Sharon Yates [i]Originally posted by dlee92:[/i][quote]Hello there, I have been using the QUINT workflow to count active neurons and have noted that when a single neuron (approximately 1-3 pixels) in a region, that the Nutil Quant report output will report 0 pixels. Could someone point me in the right direction and/or explain what is going on here?\ Best, Diana[/quote]</blockquote>

Nutil Region Area

busra — Sat, 09 Dec 2023 20:57:10 GMT

Hi, Sharon, First of all, thank you and your team for this wonderful workflow. My PhD thesis is a morphometric study on rats with Alzheimer's disease and using QUINT workflow in my study was very useful. However, I would like to ask you about a few things that are on my mind. Can I accept the Reference Region Area information given in the Nutil output report as a morphometric area? My goal is to get volumetric results from this data. For this, I will use the stereology method in my study, but I wonder if this is correct.  For example, if we talk according to the table below; can I consider the cortex area as 4772968 pixels? Also, even if I convert units from pixels to mm2 during nutil quantifier, the result is still 4772968. What should I do in this case?
Region name    Region pixels    Region area    Area unit Cortex    4772968    4.772.967.600.046    pixels Fibretracts    1755011    1.755.010.611.384    pixels Hippo    1031737    1.031.736.514.834    pixels Olfactory    3649652    3.649.652.441.756    pixels
Best regards..

RE: Combine point cloud coordinates into 1 JSON file

Britt van de Gevel — Wed, 05 Jul 2023 10:56:07 GMT

Hi Gergely, 
Thank you! I will try that. 
Best wishes, 
Britt

RE: Combine point cloud coordinates into 1 JSON file

Gergely Csucs — Wed, 05 Jul 2023 10:35:28 GMT

Hi,
There is a somewhat clumsy, but working thing at https://www.nesys.uio.no/MeshView/CloudMerger.html
It allows you loading multiple point cloud files, and then saving them combined. The button for saving is under the table.
The basic function is just loading every cloud, and putting them after each other.
One step further is when you know you don't need something, you can just remove the checkbox on the left, and that part won't be in the saved file.
And in the most tedious use case you can combine multiple clouds from multiple source files into one, via marking multiple checkboxes in a row. Names and colors are also editable.
I hope this helps, best regards,
Gergely

Combine point cloud coordinates into 1 JSON file

Britt van de Gevel — Wed, 05 Jul 2023 10:15:22 GMT

Hello, 
Due to the large size of my data I have found out I need to run the Nutil quantifier in seperate chunks. 
Does anybody know if, and how, I could combine the point cloud coordinates from these seperate chunks to get one JSON file to use in Meshview?
Thank you! 
Britt