Example datasets

The CellMorph dataset

About

The CellMorph dataset is a publicly available dataset that supplements the paper Clustering phenotype populations by genome-wide RNAi and multiparametric imaging by Florian Fuchs, Gregoire Pau, Dominique Kranz, Oleg Sklyar, Christoph Budjan, Sandra Steinbrink, Thomas Horn, Angelika Pedal, Wolfgang Huber and Michael Boutros, published in Molecular Systems Biology, June 2010.

For more details on this paper, visit http://www.ebi.ac.uk/huber-srv/cellmorph/index.php

Download

A subset of this dataset has been prepared to serve as an example dataset, to demonstrate how Phaedra protocols are set up and how imports are performed.

This prepared dataset is 10GB in size, and can be downloaded here: cellmorph.zip

It contains, for a single plate readout named 001-01, the following data:

  • Source images (16bit grayscale TIFF), one file per well per channel. E.g. Well B002.FITC.tif
  • Segmentation overlays (8bit grayscale TIFF), one file per well. E.g. 001-01-B02_segmask.tif
  • Measured cell data (tab-separated TXT), one file per well. E.g. 001-01-B02_ftrs.tab

Some additional notes:

  • The images consist of multiple fields that are already stitched together in a rectangular mosaic.
  • The folder structure has been arranged in a convenient format for Phaedra: a single folder per plate containing all data pertaining to that plate.
  • The zip file also contains a file called cellmorph protocol template.txt. This is a protocol template, to be used when setting up the protocol (see next topic).

How to Import

To import the dataset in a Phaedra environment, follow the steps outlined below.

  1. Download and unzip the dataset to a location on your computer
  2. Start Phaedra and log in to your environment
  3. Select File > Create new Protocol
  4. Select Template Wizard
  5. Select the Browse... button and select the file cellmorph protocol template.txt in the dataset
  6. Select Finish

cellmorph1.png

If the protocol creation succeeds, you will get a message mentioning the name and ID of the new protocol. Now to import the dataset.

  1. Select Tools > Import Wizard. Make sure the two Prompt to create undefined features settings are selected.
  2. Select the Generic Importer
  3. In the first page (Select Import Type), just select Next
  4. In the second page (Select Source Folder), enter the path of the unzipped cellmorph folder and select Next
  5. In the third page (Select Experiment), select the CellMorph Demo protocol, then select the New... button to create a new experiment to import the dataset into
  6. Select Finish

The import will start and, depending on your hardware, take between 5 and 15 minutes to complete.

cellmorph2.png

Configuring the Protocol

With the protocol set up and the dataset imported, we could start exploring the data now. But the protocol's configuration is still fairly minimal. Let's improve the configuration a bit so we can access more advanced functionality later on.

Note: while we talk about configuring the protocol, we are actually editing the protocol class, which can be seen as the 'blueprint' of the protocol, and which contains all the configuration settings.

To edit a protocol class, follow these steps:

  1. In the Navigator view, open the Protocols section and double-click All Protocols
  2. In the list of protocols that appears, right-click CellMorph Demo and select Edit Protocol Class

Editing a protocol class can be a complex undertaking, as there are literally hundreds of settings to adjust. Here, we will focus on two things:

  1. Attach a calculation formula to the CellCount feature, so we can see a heatmap with the cell count of each well.
  2. Configure the 'cell position features', so Phaedra can calculate cell positions, allowing us to select cells directly on the well image.

Attach a calculation formula to the CellCount feature

  1. To show the list of well features, select the Well Features tab at the bottom of the page
  2. There should be two features listed after doing the import: CellCount and WellId. Select CellCount
  3. In the panel on the right, select the Calculation section cellmorph3.png
  4. Make sure the JEP language is selected, then select the Edit expression... button
  5. In the expression dialog, select the count(x) function, then select the int subwell feature. The formula should now look like this: cellmorph4.png
  6. Select Save to apply the formula and close the dialog

Note: actually you can select any subwell feature, because this formula only needs to know the number of values, not the values themselves.

Configure the 'cell position features'

  1. Select the Sub-Well Features tab at the bottom of the page
  2. Select the x feature
  3. In the panel on the right, open the Position role list and select Center X
  4. Repeat these steps for the y feature and select Center Y for its position role
  5. Close the editor by clicking the close button on the tab at the top of the page. Select Yes when asked to save the changes

cellmorph5.png

Next steps

With the dataset imported and the protocol configured, you can now access a huge set of tools to inspect, calculate and validate the dataset. Take a look at the Application Tour to learn more about these functionalities!