Purpose and Intended Audience
Phaedra is an application that imports life-science data from lab instruments or image analysis software, and processes this data into normalized numeric values and graphical representations, such as dose-response curves. During this process, Phaedra offers a set of interactive tools for inspection, correction and quality assurance of the data:
- Inspection: using visual tools including heatmaps, charts, tables, image viewers, etc.
- Correction: using methods such as data point rejection and cell classification
- Quality Assurance: using robustness statistics, validation and approval flags, and modification history
Phaedra is designed for use by scientists performing high content screening assays:
- Laboratory scientists: preparing the plates and performing the readouts
- Chemists, biologists, statisticians: studying the reported compound effectiveness
- Managers: verifying and approving the result data
After installing Phaedra, a new icon will appear on the desktop:
Double-clicking on this icon will launch the Phaedra application. First, you will be prompted with a login screen:
When connecting to an embedded, single-user environment, authentication is not required. For distributed environments, you must authenticate using the credentials your Phaedra administrator has provided.
Updating the Application
Phaedra is a constantly-evolving application, with new features and bug fixes being released on a regular basis. To make sure your application stays current, an automatic update check is performed every time you start Phaedra. If an update is available, you will receive a notification pop-up in the lower right corner of the screen, as seen below.
Click on the 'Updates Available' title to perform the update.
Alternatively, you can check for updates manually, by going to the main menu and clicking Help > Check for Updates:
When an update is available, a dialog with the updates will appear. Press "Finish" to install the updates.
When the update has completed, You must restart Phaedra for the installation changes to take effect:
At any time, you can press the F1 key in Phaedra to open a Help screen. This screen will attempt to display help on the view currently displayed in Phaedra. This is called context-sensitive help. If Phaedra is unable to show context-sensitive help, it will display an index of help topics instead.
If you want to request access for Phaedra, or you encounter a problem in Phaedra the Help system does not have a solution for, please contact your Phaedra administrator.
But first, make sure you are using the latest version of Phaedra. To verify this, click on Help > Check for Updates:
The workbench is the window that appears after logging in to Phaedra.
At the top, just below the title, is the Menu bar. Below the menu bar is the Toolbar, containing buttons to perform several common tasks.
Below the toolbar is the main area, where a number of views can be displayed. By default, you will see two views: the Navigator and the Feature Selector. When new views are opened, they are usually displayed in the right side of the application window.
At any time, you can resize a view by clicking and holding its border, and moving it to the desired size. Alternatively, you can click and hold the view's title tab to position it elsewhere in the workbench.
If the workbench becomes too crowded, you can return to the original layout by opening the Window menu in the menu bar, and selecting Reset Perspective.
While navigating through data in Phaedra, you will notice a number of views are placed on top of each other. At the top, the view titles are arranged like stacked tabs. This represents your navigation history: you can return to a previously visited object or list by selecting the corresponding tab.
You can close a tab by clicking on the cross button to the right of the tab's title. Alternatively, you can close all stacked tabs by right-clicking on any tab and selecting Close All.
This topic explains how to navigate in Phaedra, finding your experiments and plates, and using drill-down navigation to see the amount of detail you want to see.
The starting point for navigation is the Navigator view. This view is open by default, in the top left of the Phaedra workbench.
The Navigator contains several folders. Right now, we want to access the item My Experiments under My Team. Double-clicking on this item will list all the experiments that you have imported yourself.
If you want to look at a colleague's plates, use Team Experiments instead.
If you double-click on an experiment, its plates will be listed in a new view.
This method of double-clicking on an item to see its contents, is called drill-down navigation, and is a method used frequently in Phaedra.
Note that the previous view, My Experiments, is still listed in the stacked tabs at the top of the view. You can go back to it by clicking on it, or by closing the current view.
To open a plate, double-click on it.
When opening a plate, you will be presented with a heatmap view of the plate, with additional information shown on top of the heatmap.
By default, two labels are shown on each well: the feature value and the well type. The heatmap color also represents the feature value, between blue (the plate's minimum), white (the plate's average) and red (the plate's maximum).
To know which feature is being shown, and to change it, we need to look at the Feature Selector view. This view is positioned left, below the Navigator.
The Active Feature is the feature that is being shown in the heatmap. Below it, the feature's Active Normalization is shown, as well as a preview of the coloring method used by the heatmap.
You can freely change these selections at any time; they only affect what's being shown in other views of Phaedra.
In order to view and work with data in Phaedra, you must import that data first. This means you must provide Phaedra with:
- A protocol, which specifies the list of features and image channels you're going to work with, and how they should be obtained from the source of data (see below).
- The source of data, which is usually a folder containing data and image files. This data should be structured and formatted in a way Phaedra can understand and parse.
In a typical HCS assay, the assay development process may involve creating several protocols, adjusting them as you go by adding or removing features, trying different stains and markers, etc. Eventually, a stable protocol will be created to perform the primary screen.
A protocol may contain any number of experiments, as long as the data inside them adhere to the same rules regarding features and image channels. For example, if you decide to switch from a DAPI stain to a Hoechst stain, you should create a new protocol, to:
- Formally indicate this change
- Be able to import features that may be named differently, e.g. Cell-Mean-DAPI-Intensity -> Cell-Mean-Hoechst-Intensity
- Be able to configure your image channel differently (Hoechst may have a different contrast range vs DAPI)
Creating a new Protocol
To create a new protocol, go to File > Create New Protocol. A dialog will open, listing the available protocol wizards. Select the wizard that corresponds to the instrument or software you are using, or select the Template Wizard if none matches your setup.
Note: New wizards are being added on a regular basis. If you don't see a wizard for your setup, head over to the Support section and describe the type of wizard you'd like to have.
Creating a Protocol Template
Note: This section only applies if you're using the Template Wizard. If you are using one of the other wizards, you can skip this section and follow the wizard's instructions instead.
A protocol template is a small blueprint that specifies the key settings for a protocol. By using the Template Wizard, you can either generate a new template, or use an existing template, and then create a protocol from that template.
Most of these settings are concerned with the data file names and locations. The screenshot above shows the simple template. It makes the following assumptions:
- All data is grouped into plate folders. Each plate is represented by one folder, and the name of that folder is used as the plate's name.
- Inside each plate folder, there is a welldata file containing the aggregated well data for that plate. This file should contain one value per feature per well.
- Optionally, inside each plate folder, there is a collection of subwelldata files containing cell-level data for each well. One file per well is expected. The file should contain one value per feature per cell.
- Optionally, inside each plate folder, there is a collection of image files. Each file corresponds to one channel (or mode, stain, outline, ...) for one well.
Note: If your data or images are aggregated to the field-level instead of the well-level, you will not be able to use the simple template. Please head over to the Support section to get pointers on how to create a protocol that can aggregate field-level data.
Note: The file and folder name patterns are specified via regular expressions. If you are not familiar with these expressions, check the How To > Create a new protocol section for more information about this concept.
Running the Import Wizard
Once you have created a protocol using the Protocol Wizard, you can start importing data into it using the Import Wizard. Typically, each import will create one experiment in the protocol, containing one or more plates.
To learn how to use the Import Wizard, please consult the section How To > Import plates.
As mentioned in the Navigation Basics topic, double-clicking on a plate will open a heatmap of that plate. This is the most common type of visualization for a plate in Phaedra. In this topic, several other methods of viewing plates are explained besides the heatmap.
The heatmap is a powerful tool, as it can display a lot of information in one view. By default, a heatmap will show:
- The active feature, both as a numeric value and a heatmap color for each well
- The well type of each well
- The rejection status of each well, as a yellow (technical outlier) or red (biological outlier) cross.
Heatmaps will sometimes display grey wells, with a NaN feature value, as shown in well H12 in the image above. This means that the feature value for that well is NaN (Not a Number). This may have different causes:
- The normalized value cannot be calculated, for example because the plate has no layout linked to it yet.
- The active feature is a calculated feature whose formula contains an error.
- The well has no raw feature value, for example because the image analysis failed on that particular well image.
When opening a plate, the plate view lists two tabs at the bottom: Plate View and Table View. The Plate View contains a heatmap, as explained above. The Table View contains a table listing multiple properties and feature values for each well.
Click on the Table View tab to switch from the heatmap presentation to the table presentation.
The table contains one row for each well. The first column is a thumbnail of the well image. Note that you can make the thumbnail bigger just by resizing the column width.
Besides several well properties such as well number, well type, compound and concentration, additional columns will display the different well features.
You can sort these columns by clicking on the column title, or filter rows by typing a value in the field below the column title.
Phaedra offers a wide range of plots, from one-dimensional histograms to three-dimensional scatter plots. Most of these charts can be opened by right-clicking on a plate or well, and selecting the desired chart type in the Charts submenu.
Charts have many options and settings; you will find a detailed listing in the Howto section. The most important setting for any chart is the axis feature selection. For example, in a two-dimensional well scatter plot, the X and Y buttons in the top right of the view allow you to specify two well features to plot against each other.
Well images can be shown at any time by right-clicking on a well (in a heatmap, a table or a chart) and selecting Show Well Image.
The Well Image view will open, showing the selected well's image.
Just like the Chart views, this view has many additional settings, which are explained in a help page you can open by pressing F1. Listed below are the most important functions:
- You can zoom in and out using the scroll wheel on your mouse.
- You can pan the image by clicking and holding the left mouse button, and dragging the cursor around.
- By default, all image channels are displayed on top of each other. You can toggle channels on and off using the channel buttons directly above the image.
If the protocol has been configured to calculate dose-response curves, you can view those curves in two ways:
- Right-click on a plate and select Browse Dose-response Curves
- Right-click on a well and select Show Dose-response Curve
When using the first method, a new view will open listing all the compounds in the selected plate, along with the dose-response curves for all features that have a curve configured for them.
When using the second method, a smaller view will open to the right of the screen, showing a single dose-response curve in greater detail.
In the image above, the blue curve represents the fitted dose-response curve. The dotted black line indicates the pIC50 value, while the dotted red and green lines represent the lower and upper bounds respectively. The blue circles are the well samples that were used for fitting the curve.