Analysis Simplified
Decisions based on data drive faster discovery. To speed up discovery we need a High Content Analysis solution with an intuitive, easy-to-navigate User Interface (UI) that gives scientists confidence in their data analysis.
How sensitive is your analysis to outliers, artifacts, and subjective parameterization? Clairvoyance™ analysis solution provides data visualizations and purpose-built QC methods to support scientists throughout protocol optimization and ensure they proceed in the right direction. Clairvoyance is designed to maximize data quality across all aspects of High Content Analysis. Using machine learning and deep neural net-based artificial intelligence (AI) Clairvoyance removes variability and delivers the confidence you need to make the right decisions.
In combination with the speed of Araceli Endeavor®, Clairvoyance™ creates a complete platform solution for High Content Imaging.
Dive into Clairvoyance™ Through User-Centered Video Tours
Explore how Clairvoyance™ empowers users of all experience levels with an intuitive workflow and visually guided interface. These demonstration videos guide you from system setup to data export, no prior imaging analysis expertise required.
Step-by-step tutorial on setting up your Clairvoyance software account, including login, password creation, and license agreement. 0:41
Transcript
00:01 – 00:07
When a user launches Clairvoyance for the first time, they will be prompted to log in.
They can use their email or username, depending on what was used to create the account.
00:07 – 00:15
They will also need the temporary password that was set at registration.
Next, they will be prompted to create their own password.
00:22 – 00:27
And finally, before they can start using Clairvoyance, they will have to accept the End User License Agreement.
Navigating The Home Screen
Learn how to sort and filter plates and protocols on the Clairvoyance home screen. Use toggles, formats, and search tools to streamline your workflow. 0:31
Transcript
00:00 – 00:05
The controls at the top of the home screen let you sort and filter through the plates and protocols in the system.
00:07 – 00:12
You can use the search bar to filter by name, and you can use the toggles on the right to control what kind of item you're viewing.
00:14 – 00:20
You can also choose in what format — card or list — and the number of items per page.
Image Viewer Controls
Learn how to adjust image display in Clairvoyance using FOV viewer controls. Includes grayscale, pseudo color, channel scaling, and flat field correction. 0:46
Transcript
00:00 – 00:06
Above the FOW viewer, you will find controls to let you adjust what you're seeing across the different image viewers.
00:08 – 00:13
You can see the images in grayscale or in pseudo color. You can also control what you can see across the channels in terms of intensity.
00:18 – 00:27
There are settings that let you auto per channel, and also auto all channels with a scaling method of your choice.
00:30 – 00:37
Additionally, you can view pre-processed images for plates where there has been flat-field image correction done.
Managing Task Queue
Learn how to manage running analyses in Clairvoyance. Cancel, reorder, or prioritize jobs in the task queue to control workflow execution. 0:46
Transcript
00:00 – 00:05
So I have a job running. You notice that there's a queue here. Any time I want to add a new analysis here, it's going to be added to the queue.
00:11 – 00:19
I can then go, if I realize I made the mistake and don't want to do this anymore, or want to prioritize the job that I just added. I can cancel things on the queue.
00:22 – 00:36
And that means that basically now what you'll see is going to be after this is complete. The job I just wanted is going to run, as opposed to the several jobs that I had in the queue after that.
Plate Archive vs. Removal
Learn the difference between archiving and permanently removing plates in Clairvoyance. Preserve analysis data or clean your workspace effectively. 0:58
Transcript
00:01 – 00:04
For managing plates that are not currently being used in Clairvoyance, there are two options. The more temporary way to remove a plate is to archive a plate.
00:08 – 00:16
This removes it from the UI, but does not destroy any data associated with the plate, including any analysis results. This plate can be re-imported into the system, and all information associated with it will be reinstated.
00:25 – 00:33
A more permanent action is remove plate. This will remove all information associated with the plate, including any analysis results.
00:35 – 00:41
When you click remove plate, you will see a pop-up which lets you know what exactly this will do.
00:41 – 00:49
From here, you can continue and permanently remove the plate and all its previous results, or you can cancel and back out.
Analyze Plate Data
Learn how to select a plate and review protocol information in the Analyze tab of Clairvoyance. Supports plate search, protocol lookup, and data validation. 0:43
Transcript
00:00 – 00:05
Now that we've measured our data, let's go and analyze it. In this analyze tab, we first select the plate.
00:07 – 00:13
Fortunately, this is searchable, because I have a couple hundred plates in here. And so I can pull up that plate that we just looked at.
00:15 – 00:20
Select that, and with that selected plate, you have things like a plate map. Here you can select exactly which protocol you want to look at.
00:24 – 00:31
And you can do things such as look at that protocol and make sure it's the one that you think it is. Again, this is read-only, so you know exactly what you did here.
Explore Images
Learn how to use the Measurement Tool, Scale Bar, and Pixel Spy in Clairvoyance. Measure distances, view scale dynamically, and inspect pixel intensity values. 1:23
Transcript
00:01 – 00:07
Below the FOV viewer, there are a few tools that can be helpful when exploring images in Clairvoyance.
00:07 – 00:12
The measurement tool lets you measure distance between two points in an image. To use this, click once to toggle it on.
00:15 – 00:22
Click on the image to set the first point, drag your mouse across, and then click once again to set the second point.
00:22 – 00:33
The little drop-down arrow beside this tool lets you set the unit that you want to see the distance in, as well as the color of the line for best contrast.
00:37 – 00:42
There is also the scale bar tool, which is enabled by default. You can turn it off if you'd like.
00:46 – 00:56
You can also set it to a dynamic range, which is calculated by Clairvoyance automatically, or a specified range that you can control.
00:59 – 01:15
In addition, there is a Pixel Spy that appears on the bottom right of the FOV viewer. Here you can see values for the channel intensities, as well as the x and y coordinate for any given point of the image where your mouse is hovering over.
Importing Plates and Protocols
Learn how to import plates and protocols into Clairvoyance. Includes selecting image directories and reviewing plate metadata after upload. 0:50
Transcript
00:00 – 00:05
After you've launched Clairvoyance and successfully logged in,
you will be brought to the home screen of the application.
00:05 – 00:11
Initially, the screen may mostly be blank for you,
barring some of the controls at the top.
00:13 – 00:18
But as you start to import plates and add or import protocols into the system,
they will show up as cards with metadata.
00:18 – 00:25
As you can see here on this screen, to import the plate,
click on the import button and navigate to the directory that has the images.
00:29 – 00:35
Once you've done that, you will notice that a new card will appear,
with metadata about the plate that you've just imported.
Opening the Process Screen
Quickly access the Process screen for any plate in Clairvoyance. Use the plate shortcut icon or select a plate from the dropdown menu on the home screen. 0:35
Transcript
00:01 – 00:05
To process an imported plate, you have two options. You can click on the process icon, which is the bottom-left icon on every plate.
00:05 – 00:12
This will bring you to the process screen, with the plate already loaded. Alternatively, you can, from the home screen, navigate to the process screen.
00:18 – 00:24
From this dropdown, select the plate that you'd like to process.
View Pre-Processed Images
Learn how to use the pre-processed image toggle in Clairvoyance to view images corrected for background or illumination, using flat field calibration data. 0:26
Transcript
00:01 – 00:07
In addition, the view pre-processed image toggle can be turned on to view images that have been corrected. These corrections apply to background or illumination.
00:11 – 00:16
Note that this is only available on plates which have flat-field images associated with them.
Exporting Results Part 1
Learn how to export summary and object-level measurement data in Clairvoyance as CSV files. Includes per-well, per-FOV, and auto-export options. 1:48
Transcript
00:01 – 00:08
Now that my analysis is finished, I can go and view this data. The first way to do that is through heat map on the process screen.
00:08 – 00:19
So here it's displaying a particular measurement. Any given measurement I can choose to display, as well as look at a given metric for that measurement.
00:19 – 00:24
Similarly, so say I want to care about spots per nuclei. It will show that, and you can also export those results.
00:24 – 00:34
This export has a couple different options. You can choose how the naming convention for those, as well as aggregation on the summary CSV file.
00:37 – 00:41
This will be one line per well versus one line per field of view. You also choose whether or not you want to export object summaries.
00:44 – 00:51
These are summaries that include for each object. There will be a line on the CSV and include the measurements that are specified.
00:56 – 01:02
If I don't care about nuclei, I won't export those. But this will give me a per spot look at my data.
01:02 – 01:06
Again, I can choose to define what measurements they are. When I press export, I'll be prompted to put a location for this.
01:11 – 01:16
I'll select a particular folder, and it will automatically export those. If I wanted to do that automatically, I can do that as well.
01:20 – 01:28
And this I would check this auto export results. Similarly, I have settings for this, and I can choose what results I want auto exported.
01:28 – 01:37
With that auto export, whenever I press run, it will automatically spit out that CSV. And notice that I can define my directory to auto export here.
Exporting Results Part 2
See how to access exported CSV results from Clairvoyance and bring them into external tools like Prism or continue analysis inside Clairvoyance. 0:46
Transcript
00:02 – 00:07
Now let's go and look at the CSV files that were created. I saved it to desktop, so I’ll minimize the screen, and here they are.
00:09 – 00:15
The object file for every spot, as well as the summary file. Looking at the summary file, we'll see that I get all 24 wells that I have analyzed.
00:20 – 00:26
And we have our metadata here. And then we have quite a bit of measurements here.
00:26 – 00:31
So you can either put this in your favorite analysis program, such as Prism. Or you could stay within our ecosystem.
00:33 – 00:38
And in the next video, I'll show you how to analyze data within Clairvoyance.
Understanding Notifications
Learn how to use the notification panel in Clairvoyance. Understand message types like info, success, failure, and partial success, and view system logs directly. 1:10
Transcript
00:00 – 00:05
The notifications panel can be opened by clicking on the bell icon on the bottom left in the status bar.
00:08 – 00:14
This panel shows different kinds of notifications triggered by actions taken in Clairvoyance. Clairvoyance shows four different kinds of notifications.
00:18 – 00:29
Informational, represented by the blue icon. Successes, represented by the green icon. Failures, indicated by the red icon. And finally, the partial successes, indicated by the yellow icon.
00:32 – 00:39
The notifications panel can show up to 100 of the most recent notifications at any given time.
00:39 – 00:46
To access a full list of system logs, click on the System Logs link at the top of the notifications panel.
00:46 – 00:51
Alternatively, this can also be accessed by the menu on the top right.
00:51 – 00:58
Once in the system log page, you can download them in a zip folder or clear them.
Cytoplasm and Membrane Masking
Learn how to subtract nuclei from cell masks to define cytoplasm, and how to generate membrane collars for subcellular segmentation in Clairvoyance. 2:08
Transcript
00:01 – 00:09
Now that I've found my nuclei and my cells, I can do some mask math to find various substructures.
00:10 – 00:20
The first thing I'm going to do is use the cell mask and this nucleus mask to create a cytoplasm. That is the cell mask subtracted by the nuclear mask.
00:23 – 00:35
So to do that, I will do a find cytoplasm step. So this is going to be a subtract. Now I'll say subtract my nuclei from my cells.
00:36 – 00:44
On the same token, I can do a similar thing and find something like a membrane. So I can go membrane.
00:45 – 00:59
Here these channels really don't matter very much. Mask operation, so now I'm going to say collar. And so for that, I'm going to say use my cell collar.
00:59 – 01:08
You'll notice that any object that I defined previously is appearing in this list. And so for this, let's say I have bidirectional mask, 20 pixels on either side.
01:09 – 01:11
And now I'm going to run both of these steps.
01:16 – 01:24
So let's see how we did. So again, this cytoplasm is going to be basically my cell minus the nucleus.
01:26 – 01:33
Indeed it is. We see that our nucleus is filling in those holes quite nicely.
01:33 – 01:39
Similarly, we can see how our membrane did. And that should just trace around each cell. Indeed it did.
01:40 – 01:52
It's probably a little too thick. So if I really wanted to measure something like trafficking, I would then thin this down to maybe, say, eight pixels, which again is going to be representative of about two microns.
01:52 – 01:59
But again, this is how you do basic mask operations. So you can build and find subcellular structures with this.
Running Image QC
Learn how to run image QC in Clairvoyance by using reference image similarity scoring. Flag out-of-focus views and review QC metrics like nuclei count and dead cell ratio. 1:41
Transcript
00:01 – 00:09
In this video, I'll demonstrate an image QC. I want to QC this cell positivity protocol, and I'm going to add that step to this existing protocol.
00:12 – 00:20
As you'll notice, it always appears first in this. So this image QC is actually an image similarity algorithm, so you need to find a good reference image.
00:21 – 00:27
I'm going to use my no treatment control as my reference. So with this, I'm now going to press run, and I'm going to run my entire plate here.
00:35 – 00:48
So what you'll look for in this particular algorithm is it's going to give you a score from 0 to 1. And that score is essentially going to say how similar to the reference each individual field of view is.
00:54 – 01:02
Now that my run is finished, I can view those results. So I'm going to go to my heatmap here. So you see there's a nuclei count. You can also see that there's a green nuclei count, as well as a dead cell ratio, which again follows my dilution pattern quite nicely.
01:08 – 01:15
However, what I'm interested in is this QC score. So I'm going to look at a mean. So it looks like we did indeed flag a well. And indeed, that's because these fields of view are out of focus here.
01:21 – 01:23
And so we've correctly flagged this well.
01:23 – 01:27
You’ll also notice that these are likely colored, and that's because the morphology of the image is changing.
01:27 – 01:32
But as you'll notice, the score for this is much lower than those morphology changes.
Run Batch Analysis
Learn how to efficiently run batch processing in Clairvoyance using protocols across multiple plates. Includes validation, export, and error handling steps. 2:16
Transcript
00:00 – 00:05
Once I have a protocol I'm happy with, I can then apply that protocol to multiple plates at once using the batch function.
00:07 – 00:15
So scrolling over to the batch tab, I first select my plates. One criteria is that all these plates have to be the same plate format. Here I’m going to select 96 well.
00:17 – 00:26
I'm going to look for a particular run where I was testing quite a few different plate types and add those using the select function. I can also press select all to add all of those that match the criteria there.
00:31 – 00:43
Pressing done, I'll then select which protocol I want to measure with. I can also choose if I want to automatically export those results, again specifying where those are going as well as what is to be exported.
00:46 – 00:52
Now I produce CSV files automatically when the plates are done with the analysis. Otherwise, I will access the results with each individual plate or in the analyze tab.
00:56 – 01:06
Now I'm going to press run. And the first step is going to validate these plates to make sure that these plates meet the criteria specified in this given protocol.
01:08 – 01:14
Now that these are validated, I can now see that some of them failed validation, and those are now flagged.
01:14 – 01:27
The reason those are flagged validations, I can scroll over and it’s because there's some channels that I have specified to measure that are not present in these. So I can either change my protocol or delete the plates in question.
01:30 – 01:36
Now that I have all of my plates matching the criteria protocol, I can then run these plates.
01:40 – 01:44
You’ll notice now that I have monitors coming up here saying what my batch is doing. And so here I can monitor, okay, so the first plate is running.
01:50 – 01:58
Then I'll produce the results as queued. And you notice that I can also cancel individual plate runs within this task manager now.
02:03 – 02:06
So with that, that's how you can run a batch protocol.
Correcting Illumination
Learn how to correct uneven image illumination in Clairvoyance using flat field and vignette data. Ensure cleaner, more consistent FOV intensity for your analysis. 1:33
Transcript
00:01 – 00:05
Let's go over how to correct illumination evenness in your images. All Endeavor data come with flat field files.
00:09 – 00:15
These are automatically associated upon import, indicated by this green button. You can choose to edit or remove these flat field files if you like.
00:19 – 00:24
Or if for some reason your data do not have flat field files, you can also add them here. The green means that they're associated within the file folder.
00:30 – 00:35
There's a calibration file. In those calibration files lives this vignette data with every image.
00:38 – 00:43
When you go into the process screen, you can see that there is a view preprocess button.
00:43 – 00:50
As you can see, on the corners of this image there is indeed this characteristic intensity drop-off here. And that's simply because there's a round objective with a square field of view.
00:54 – 00:59
By clicking this view preprocessed image, you're correcting that image visually.
01:02 – 01:07
As you can see, we've restored illumination in those corners here. In order to correct this within your analysis, you have to go to preprocess image on your protocol.
01:13 – 01:18
Select that you want to preprocess, and highlight this vignette and flat field correction.
01:18 – 01:23
Now whenever you run your data, you'll be running with this corrected, flatter image.
Adding Custom Nuclear Templates
Learn how to create, resize, and upload custom nuclear templates in Clairvoyance. Use ImageJ to edit templates and manage them for precise detection workflows. 2:39
Transcript
00:00 – 00:06
In this video, I'll demonstrate how to add custom templates to Clairvoyance. Sometimes our new templates aren't quite right, so what I want to do here is show you how to find the templates and edit.
00:17 – 00:23
So the templates that come within Clairvoyance are not editable, but you can add your own. So here are all the nuclear templates.
00:25 – 00:33
If you click on one of these, you'll bring up the file folder here, and so you see that we have all the templates that are included with our software.
00:37 – 00:43
So what we can do is we can load one of these images into ImageJ. You can also use your own templates as well.
00:46 – 00:54
So that's why our template looks like an ImageJ. I want to make this particular one larger. So what I'll do is I'll go size, and then I can customize my sizing.
00:56 – 01:01
And it will automatically scale this for me to say I wanted something that was a little bit smaller than this.
01:05 – 01:10
I now can save this to a place that I will recognize. I'm going to save this as a PNG file.
01:13 – 01:20
So I'm just going to put this on my desktop, but I would recommend that you put this in a safe place that it won't get overwritten.
01:27 – 01:35
All right, so I saved my template. Now I'm going to add it to Clairvoyance. So I'll go manage, add template.
01:39 – 01:45
And I'll specify that it's a fluorescent template, and it's part of my nucleus. So I'll select my image.
01:45 – 01:53
And again, this is saved on my desktop. This is what I saved here. I'll open that, and now I can add that template.
01:55 – 02:03
And I'll say it is 110 and ten pixel nucleus.
02:06 – 02:17
Now you'll see that in those references I now have that given template. You can actually add a whole category of templates here by continuing to add templates into this given category.
02:23 – 02:30
Now you'll see that I can now select that template and then use that to find my nuclei.
Linking Objects
Learn how to link measurement objects in Clairvoyance – like spots to nuclei or nuclei to cells – to quantify morphology, relationships, and per-object statistics. 1:32
Transcript
00:00 – 00:05
Now that we've defined a few measurement objects, let's link those objects together. The purpose of linking is such that you can relate one object to the other.
00:10 – 00:17
So say you wanted to find the morphology of cells with a green nuclei. Linking allows you to do that.
00:17 – 00:24
Or say in a given nuclei, how many spots are within that nucleus. That will also be accomplished in linking.
00:24 – 00:30
Some of these things have implicit links, and those are indicated here. And so if you scroll, it'll tell you what is linked there.
00:33 – 00:38
These links are created because within the cytoplasm operation you have both cells and nuclei. So Clairvoyance already knows if you want to link those together.
00:44 – 00:50
However, if you want to define links manually, you can do that as well. So say I want to link together nuclei and spots.
00:55 – 01:02
You can define how you want to link that. So I'm actually going to link that. You can choose it based on your detection or your segmentation. I’m going to choose detection here.
01:04 – 01:09
I'm going to specify the bounding boxes must be overlapped by at least 90%.
01:10 – 01:16
Now when I analyze this data, an aggregation of spots will show up in my nuclei. And I will be able to do statistics on nuclei with different amounts of spots and spot areas.
QC Filtering and Analysis
Apply QC filters in Clairvoyance using focus scores and visual tools like heatmaps and swarm plots. Exclude out-of-focus wells and export clean data. 1:25
Transcript
00:01 – 00:06
Now that we've measured the objects in this plate, we can now go load this into our analyze tab and add plots. I'm going to add a heatmap and a swarm plot here, and look at QC in both.
00:23 – 00:27
As you see, that well F1 is flagged in both. I can then click on that, and that will be loaded in my well viewer.
00:30 – 00:36
And as you see, we can see that these have lost focus in the periphery around here. So now what I can do is I can actually filter these out from analysis.
00:41 – 00:46
So I'm going to exclude those filtered wells using my QC score. So as you notice, this point F1 is there.
00:51 – 00:55
And then I can set my QC score to around 0.6. And now all of a sudden, that well F1 is now joining the rest of its compatriots.
01:01 – 01:03
And so it looks like an entirely normal well. So now all those out-of-focus fields of view are no longer being included in that analysis.
01:10 – 01:15
And again, we can export that filtered analysis as a CSV file.
Managing the Task Queue
Learn how to manage running analyses in Clairvoyance. Cancel, reorder, or prioritize jobs in the task queue to control workflow execution. 0:46
Transcript
00:00 – 00:07
So I have a job running. You notice that there's a queue here. So any time I want to add a new analysis here, it's going to be added to the queue.
00:11 – 00:19
I can then go, if I realize I made the mistake and don't want to do this anymore, or want to prioritize the job that I just added. I can cancel things on the queue.
00:22 – 00:36
And that means that basically now what you'll see is going to be after this is complete. The job I just wanted is going to run, as opposed to the several jobs that I had in the queue after that.
Account Setup Guide
Step-by-step tutorial on setting up your Clairvoyance software account, including login, password creation, and license agreement. 0:41
Transcript
00:01 – 00:13
When a user launches Clairvoyance for the first time, they will be prompted to log in with their email or username, depending on what was used to create their account, as well as the temporary password that was set.
00:13 – 00:18
Next, they will be prompted to create their own password.
00:22 – 00:29
And finally, before they can start using Clairvoyance, they will have to accept the End User License Agreement.
Understanding Notifications
Learn how to use the notification panel in Clairvoyance. Understand message types like info, success, failure, and partial success, and view system logs directly. 1:10
Transcript
00:00 – 00:05
The notifications panel can be opened by clicking on the bell icon on the bottom left in the status bar.
00:08 – 00:14
This panel shows different kinds of notifications triggered by actions taken in Clairvoyance. Clairvoyance shows four different kinds of notifications.
00:18 – 00:29
Informational, represented by the blue icon. Successes, represented by the green icon. Failures, indicated by the red icon. And finally, the partial successes, indicated by the yellow icon.
00:32 – 00:37
The notifications panel can show up to 100 of the most recent notifications at any given time.
00:39 – 00:46
To access a full list of system logs, click on the System Logs link at the top of the notifications panel.
00:46 – 00:51
Alternatively, this can also be accessed by the menu on the top right.
00:51 – 00:58
Once in the system log page, you can download them in a zip folder or clear them.
Creating and Loading Protocols
Learn how to load existing protocols or create new ones in Clairvoyance. Define measurement objects like nuclei, spots, and cytoplasm, and run steps sequentially. 2:03
Transcript
00:00 – 00:06
In this video, I will demonstrate how to load a protocol as well as create a new protocol. To load a protocol from the process screen, you can select this protocol tab, and here you'll find a dropdown with all of the previously saved protocols.
00:16 – 00:28
This is a Genotoxicity screen, so I will select a Genotox protocol. You'll see that then there is the name of the protocol here, as well as any notes.
00:28 – 00:35
Then you'll see predefined objects that I saved as a protocol, including a find nuclei step, find spot steps, cell steps, as well as creating cytoplasm steps and ratios as well.
00:39 – 00:43
I can then run this protocol if I wished. I can now also go create a new protocol.
00:46 – 00:56
With this new protocol, I have to give a unique name, as well as write any notes. And these all will be searchable, so it's very useful for multi-user setups here.
00:58 – 01:05
Within this protocol, I can now define my measurement objects, as well as do any image pre-processing in order to find various objects I select.
01:09 – 01:17
So I'm going to first put in a find nuclei step. You'll notice now that I can select which channel we are defining that object in, as well as in our metrics tab defining which channels we're actually doing this measurement in.
01:23 – 01:31
I can then add more steps, such as a find cell step, or a find cytoplasm step, or find spot step. And you notice they are all appearing below that first measurement object.
01:38 – 01:44
And now they will be run sequentially. So then you can use any given measurement object in the object below it.
01:46 – 01:54
Note that you cannot reorder these steps. So if you want to change the order, you have to delete the step you want to reorder and add it back.
Measuring Nuclei
Learn how to measure segmented nuclei in Clairvoyance, including intensity in specific channels and morphology features like area and shape. Save time with focused metrics. 1:47
Transcript
00:01 – 00:07
Now that I've found and segmented my nuclei, it's time to measure those nuclei. What this detection segmentation gets you is just nuclear counts and their XY location in space.
00:14 – 00:19
If you wanted to measure intensity, you'd have to check our intensity metric here, which then asks you to specify which channels you want to measure intensity in.
00:25 – 00:35
For this particular assay, I want to measure channel two, my green channel, because it represents my gamma H2AX. We also have morphology, so if I want to measure morphology, I’d also check that box.
00:40 – 00:48
Something that's important to note is that we have quite a list of different morphology and intensity measurements. If I just wanted mean intensity, I would just select mean intensity there.
00:54 – 01:00
And this will be slightly faster because we're making fewer measurements. Similarly, for morphology there's quite a bit of measurements we have there.
01:03 – 01:08
And a lot of them are doing particular sorts of math on these data. And each time it does that for each object, it takes time.
01:12 – 01:18
So you can save a lot of time if you're only interested in something like the area of those nuclei or perhaps their shape, and you could just select those corresponding measurements.
01:26 – 01:30
All of these measurements are further described in our wiki as well. So once I've selected those measurements, I can now go and I have to rerun these data in order to get those measurements there.
Linking Objects
Learn how to link measurement objects in Clairvoyance – like spots to nuclei or nuclei to cells – to quantify morphology, relationships, and per-object statistics. 1:32
Transcript
00:00 – 00:05
Now that we've defined a few measurement objects, let's link those objects together. The purpose of linking is such that you can relate one object to the other.
00:10 – 00:17
So say you wanted to find the morphology of cells with a green nuclei. Linking allows you to do that. Or say in a given nuclei, how many spots are within that nucleus. That will also be accomplished in linking.
00:24 – 00:30
Some of these things have implicit links, and those are indicated here. And so if you scroll, it'll tell you what is linked there. These links are created because within the cytoplasm operation you have both cells and nuclei. So Clairvoyance already knows if you want to link those together.
00:44 – 00:50
However, if you want to define links manually, you can do that as well. So say I want to link together nuclei and spots. You can define how you want to link that. So I'm actually going to link that.
00:59 – 01:06
You can choose it based on your detection or your segmentation. I’m going to choose detection here. I'm going to specify the bounding boxes must be overlapped by at least 90%.
01:10 – 01:21
Now when I analyze this data, an aggregation of spots will show up in my nuclei. And I will be able to do statistics on nuclei with different amounts of spots and spot areas.
Filter Measurement Results
Learn how to filter individual objects or wells in Clairvoyance by setting size thresholds. Export cleaner imaging results for high-content analysis. 1:14
Transcript
00:01 – 00:06
Now I've graphed some data. I can also filter these data. The nice thing is I can filter individual objects as well as whole wells.
00:10 – 00:15
Let's say I want to exclude nuclei that are too small. So these probably aren't real nuclei. They might be things like micronuclei. And I'm obviously measuring them, so what I can do is go ahead and set a minimum there.
00:24 – 00:28
As you can see, all of my measurement objects refresh as I'm doing this. If I set it much too high, it gives me all these strange results. But if I set it to a point that actually makes sense — so basically if you're not above around 500 pixels — you know it's probably not a real nucleus.
00:42 – 00:45
I can actually get much better, cleaner results that way. If I want to, I can then go ahead and export those filtered results. And that way I have nice clean data to look at.
00:53 – 01:01
Another fun thing you can do with these graphs is you can look at the XY coordinates as well as trace the data back to whatever you see there.
Run Batch Analysis
Learn how to run batch processing in Clairvoyance using protocols across multiple plates. Includes validation, export, and error handling steps. 2:16
Transcript
00:00 – 00:07
Once I have a protocol I'm happy with, I can then apply that protocol to multiple plates at once using the batch function. Scrolling over to the batch tab, I first select my plates.
00:11 – 00:17
One criteria is that all these plates have to be the same plate format. Here I’m going to select 96 well. I'm going to look for a particular run where I was testing quite a few different plate types and add those using the select function.
00:23 – 00:31
I can also press “Select All” to add all of those that match the criteria there. Pressing “Done,” I'll then select which protocol I want to measure with.
00:33 – 00:43
I can also choose if I want to automatically export those results, again specifying where those are going as well as what is to be exported. Now I produce CSV files automatically when the plates are done with the analysis.
00:50 – 00:56
Otherwise, I will access the results with each individual plate or in the Analyze tab. Now I'm going to press “Run.”
00:57 – 01:06
The first step is going to validate these plates to make sure that they meet the criteria specified in this given protocol.
01:08 – 01:21
Now that these are validated, I can see that some of them failed validation and those are flagged. The reason is that there are some channels I specified to measure that are not present in these plates.
01:24 – 01:30
So I can either change my protocol or delete the plates in question. Now that I have all of my plates matching the protocol criteria, I can then run these plates.
01:40 – 01:56
You’ll notice now that monitors are coming up here saying what my batch is doing. I can monitor: okay, the first plate is running, and then it will produce the results as queued. I can also cancel individual plate runs within this task manager.
02:03 – 02:06
So with that, that's how you can run a batch protocol.
Finding and Segmenting Nuclei
Learn how to detect and segment nuclei in Clairvoyance using fluorescence, template matching, neural networks, and threshold segmentation tools for accurate analysis. 6:32
Transcript
00:01 – 00:09
In this video I will demonstrate how to find and segment nuclei. For this, we'll add the “Find Nuclei” step. First, specify the channel. In this channel, our host staining is in blue, so that will be in channel one.
00:17 – 00:28
The fastest way to find nuclei is through template matching. You can also use a histogram-based methodology, local peaks, or a deep neural net. The deep neural net will be slower, but more accurate, particularly with dense nuclei.
00:35 – 00:53
For the template matching step, you first choose a reference. We have references based on fluorescent nuclei, as well as various transmitted light references (focused, over-focused, and under-focused templates). These depend on the state of your nuclei. For this example, we're going to focus on fluorescence.
01:00 – 01:22
Once selected, I then choose the size of my nucleus. I'm going to use a 100 × 100 pixel size, as we previously determined using the measurement tool. Once I know the size of those nuclei, I can say where I want to measure this. If a mask was previously defined, I can use it. Right now, since this is the first step, I can only say “within image.”
01:32 – 01:58
With that, I have advanced options. First is luminance, so I can use Pixel Spy to determine the luminance of background versus foreground. My background maxes at about 300 A.U., so I can say nuclei are anything above 400. We also set a contrast threshold—this is optional but can filter results. Unchecking it speeds things up slightly.
02:07 – 02:38
The next option is a complexity threshold, basically how close the nucleus is to the template. Less round or irregular nuclei need a lower confidence threshold. Too high may produce false positives. The final setting handles spacing—how close nuclei are to each other and how much overlap is tolerated. I’ll leave these defaults and run this step.
02:57 – 03:31
The output is bounding boxes, which define the XY centers. These boxes look too small and seem to double count, so I’ll increase the reference size and be less tolerant of overlap. Running again shows better results.
03:39 – 04:05
Now onto segmentation. The difference: detection finds only the XYZ center and bounding boxes, while segmentation defines the boundaries. As with detection, there are multiple methods: deep neural net (AI models), threshold-based, and watershed (threshold + edge detection). We’ll use threshold-based here.
04:39 – 05:13
Threshold options include: automatic (mean/median/max), manual (user enters boundary), Otsu/histogram methods. These can be applied at the field-of-view level or single-object level. Field-level is best for uniform intensity; object-level helps capture dim objects. I’ll bias the mask slightly to make it less restrictive and run.
05:35 – 05:58
The run button executes the protocol at different levels: single field of view, single well, or entire plate. For this example, I’ll run it on a single field of view.
06:06 – 06:20
The result looks over-segmented, so I’ll adjust the bias slider to make it less restrictive. Now I have properly segmented nuclei.
Create Ratio Measurements
Learn how to generate ratio-based bioscience measurements like spots per nucleus in Clairvoyance. Includes spot count, object linking, and well-based analysis. 0:51
Transcript
00:00 – 00:05
Now that I have nuclei in spots, I'm going to go to the primary measurement outcome of this, which is spots within a nucleus. So I'm going to then create a ratio and call it spots per nuclei.
00:07 – 00:10
For my numerator, I have to do a spot count and then ratio that to a nuclear count. As you see, any measurement I've previously defined I can now ratio in here.
00:13 – 00:17
For this, I'm actually going to apply this analysis to an entire drug treatment as defined by this plate map. Now I'm going to press run on these selected wells.
Analyze Plate
Learn how to select a plate and review protocol information in the Analyze tab of Clairvoyance. Supports plate search, protocol lookup, and data validation. 0:43
Transcript
00:00 – 00:05
Now that we've measured our data, let's go and analyze it. In this analyze tab, we first select the plate.
00:05 – 00:09
Fortunately this is searchable, because I have a couple hundred plates in here. And so I can pull up that plate that we just looked at and select it.
00:14 – 00:18
With the selected plate, you have things like a plate map. Here you can select exactly which protocol you want to look at.
00:22 – 00:28
You can view the protocol and make sure it's the one you think it is. Again, this is read only, so you know exactly what you did here.
Cytoplasm and Membrane Masking
Learn how to subtract nuclei from cell masks to define cytoplasm, and how to generate membrane collars for subcellular segmentation in Clairvoyance. 2:08
Transcript
00:00 – 00:05
Now that I have nuclei in spots, I'm going to go to the primary measurement outcome of this, which is spots within a nucleus. So I'm going to then create a ratio and call it spots per nuclei.
00:07 – 00:10
For my numerator, I have to do a spot count and then ratio that to a nuclear count. As you see, any measurement I've previously defined I can now ratio in here.
00:13 – 00:17
For this, I'm actually going to apply this analysis to an entire drug treatment as defined by this plate map. Now I'm going to press run on these selected wells.
Analyze Plate
Learn how to select a plate and review protocol information in the Analyze tab of Clairvoyance. Supports plate search, protocol lookup, and data validation. 0:43
Transcript
00:00 – 00:04
Now that we've measured our data, let's go and analyze it. In this analyze tab, we first select the plate.
00:07 – 00:13
Fortunately this is searchable because I have a couple hundred plates in here. And so I can pull up that plate that we just looked at and select that.
00:17 – 00:22
With that selected plate, you have things like a plate map. Here you can select exactly which protocol you want to look at.
00:24 – 00:31
You can also look at that protocol and make sure it's the one that you think it is. Again, this is read only so you know exactly what you did here.
Genotox Protocol Setup Part 1
Full demo of genotoxicity assay setup in Clairvoyance. Learn how to configure protocols, segment nuclei and puncta, and normalize γ-H2AX and micronuclei counts by cell. 7:37
Transcript
00:00 – 00:06
In this video I'm going to do an end-to-end demo looking at genotoxicity analysis. First I'm going to search for my assay, which is right here.
00:11 – 00:19
You'll notice that we have both a flat field images directory to correct vignetting, as well as a plate map loaded on this. I'm going to then go into my process screen and load a protocol.
00:23 – 00:30
This particular protocol is a genotoxicity protocol that has quite a few different steps in it. As this is an older plate, I don't have an image viewer available, though new versions of Endeavor software will produce this.
00:38 – 00:47
I do have a plate map, and you can see I have four different drugs in a dilution series with replicates per row. Here I'm going to look at etoposide, a DNA damage inducer.
00:52 – 00:59
Etoposide induces both double stranded breaks as well as micronuclei. I'm going to press run for this well in this protocol to demonstrate what each step does.
01:09 – 01:20
The primary outcomes of this assay are double stranded breaks as marked by gamma H2AX, and micronuclei in the nuclear channel, which here is a Hoechst stain. These are chromosomal fragments that have broken off and been expelled by the cell.
01:28 – 01:37
We also have a cell mask stain to segment the cytoplasm. Together, we'll use all three channels to identify nuclei, find gamma H2AX puncta marking double stranded breaks, detect cells, and make a cytoplasmic mask.
01:47 – 01:59
Then we’ll find micronuclei in the cytoplasmic mask and ratio gamma H2AX puncta as well as micronuclei on a per-nucleus basis, giving us an output normalized to cell count. My first step is looking at the Hoechst stain to mask nuclei.
02:08 – 02:23
This uses our deep neural net. I use template matching first to find cells, then the fast deep neural net comes in to refine within those squares. This is important because thresholding would segment micronuclei as nuclei, while the neural net has been trained only on nuclei.
02:40 – 02:45
This gives me both a cell count and a segmentation mask. Within this segmentation mask, I'm now going to look for my gamma H2AX puncta.
02:51 – 03:03
There are many puncta, so I’ll focus on the bright ones. Using template matching with a small template size works well, since these double strand breaks manifest as puncta about 0.5–0.75 microns in size.
03:18 – 03:25
That step did a good job, and then we segment those puncta. This allows me to measure intensity in channel two as well as morphology, with optional additional measurements if desired.
03:39 – 03:46
Now we have counts and segmentations for both puncta and nuclei. Next, we’ll use channel four to find cells, based on the detected nuclei.
03:57 – 04:07
Looking around each nucleus, we can segment the cellular area. This mask is not object-based—it only distinguishes cell versus not cell—so we can identify the cytoplasm.
04:22 – 04:33
We then subtract nuclei from the cellular mask, giving us the cytoplasm mask. This is population-based and serves as the area to look for micronuclei.
04:49 – 05:06
Within the cytoplasm mask and Hoechst channel, we look for micronuclei. These are chromosomal fragments that break off due to DNA damage, and they are much larger than gamma H2AX puncta, so we use a larger template.
05:18 – 05:31
Identification shows we can detect small fragments but miss very large breaks, since template match fails when objects exceed the template size. A larger template could be used, but overall the segmentation works well.
05:59 – 06:14
Finally, we create ratios of puncta counts to nuclei counts, normalizing results to cell count. Preprocessing includes flat field correction for intensity accuracy.
06:33 – 07:07
I also link objects—nuclei, cells, and cytoplasm are implicitly linked by use in sequential steps, and custom links can be added via bounding box overlap. This lets me summarize and associate measurements across the whole cell and its parts. Since I’m happy with this analysis, I’ll run it.
Genotox Protocol Setup Part 2
Learn how to efficiently analyze genotoxicity imaging data in Clairvoyance using dose-response curves, EC50 calculations, interactive scatter plots, and versatile export tools. 3:42
Transcript
00:01 – 00:08
Now that this complex protocol is finished running in about 23 minutes, we can go and view our results. I have this data set loaded up and can now select the analysis we just did.
00:15 – 00:22
As you can see, we have our plate map here. I can select and look at various things in the well viewer, as well as change our channels and settings.
00:26 – 00:33
I'm also going to add a few plots. The primary plot I’m going to look at is this dose response curve, where I can get a ratio of gamma H2AX per nucleus and select the treatment.
00:38 – 00:46
I’m going to look at Ametycin, and you can see this very lovely dose response curve here. It shows an R-squared of 0.93, and we can visualize any given data point to see what it looks like.
00:55 – 01:04
With that A11, I can click on it, and it will come up in my image here. I can also add another dose response curve looking at a different response, such as micronuclei.
01:09 – 01:19
For this one, I’ll look at Etoposide, which gives a 0.98 R-squared. We’re also getting automatically calculated EC50s or IC50s if it’s an inhibitory reaction, as well as the half-max.
01:25 – 01:32
We can also look at things like a box and whisker plot, or create a scatter plot. Any metric we have here can be visualized, such as area mean of micronuclei, to see the distribution.
01:43 – 01:50
For example, well A1 seems to be getting micronuclei that are larger than predicted. I can go back and look at that particular subset to confirm.
01:52 – 02:02
Going into this well, we can zoom in and see whether there are indeed large micronuclei. It looks like they are getting quite a few large nuclei there.
02:04 – 02:13
We can also create scatter plots, comparing gamma H2AX versus micronuclei. They are pretty well correlated, but you’ll notice two different trends emerging.
02:15 – 02:24
Rows E and F, which correspond to the Etoposide treatment, have more micronuclei per nucleus than gamma H2AX. Meanwhile, rows A and B with Ametycin have more gamma H2AX than micronuclei.
02:37 – 02:46
We can then look at the mechanism of action of those compounds. I can also export results as a CSV, which contains the same data shown in the graph.
02:47 – 02:56
These plots can also be exported as SVG or PNG images. I can even lasso and zoom in on particular points if needed.
02:59 – 03:06
If I want, I can go back to my process page and export the whole result set. From here, I can load my export results.
03:12 – 03:21
I’ll specify exactly which object summaries I want, such as a summary per nucleus and aggregation per well. I don’t need spot-level summaries for this export.
03:26 – 03:34
Pressing the export button will save those files onto my desktop. And with that, there’s the end and demo of our genotoxicity assay.
Capabilities
Clairvoyance delivers:
Simple Workflow - Intuitive, yet powerful UI to select, segment and review.
Wide Variety of Detection and Segmentation Methods - Optimize analysis protocols to use the best tool for the job, across all assays.
Neural Net Optimized Segmentation - Utilize trained neural networks for segmentation of cells, spots, and nuclei.
Visualization Tools - Explore and interact with data and images, and understand trends and hits using heatmaps, scatter plots, box and whisker plots and dose response curves.
ClaireRT Monitoring Software
ClaireRT is designed to streamline workflows in ultra high throughput screening, this innovative product empowers you to improve data quality and lab efficiency.
ClaireRT addresses challenges in traditional quality control methods by identifying issues before they escalate. From cell biology to staining and imaging parameters, it ensures each step meets your quality standards—all in real time.
Key Benefits:
Reduce Waste: Less time and resources spent re-running poor-quality plates
Save Costs: Avoid expensive re-plating
Boost Efficiency: Run more screens through your lab
Improve Accuracy: Achieve better analysis results
Clairvoyance Difference
Clairvoyance’s simple and powerful workflow combined with deep neural networks is designed to provide accurate data with minimal user input or training.
SIMPLICITY
In addition to image analysis algorithms for feature detection and segmentation, Clairvoyance offers neural networks to simplify the task of segmenting nuclei, cells and other subcellular objects.
NEURAL NETS
Tailor your analysis times to align seamlessly with your compute resources and assay requirements.
FLEXIBILITY
Accelerate your productivity journey with our cutting-edge analysis software and workstation – where getting results quickly is a certainty.
SPEED
Clairvoyance supports your visualization and analysis needs. Review and interact with images – from plate to well to single field of view – and measurements together to make better decisions.
VISUALIZATION