High-throughput screenings with multi-well plates
To easiest way to work with multi-well plates for high-throughput screenings is through the Plate analysis window.
A quick demonstration of the use as a video can be found here. It goes through loading the data, using an existing flow, viewing the images, extracting the parameters and showing them.
There are three data sources to keep track of:
- Images: This should be the root of an image dataset. Channels are collected from here.
- Flow: The flow script used to process the images.
- Measure values: Where the data is stored, or should be stored
You don't need to specify all data source:
- To view images, you only need to specify the images
- To process image you need to specify all sources (just create a new empty measure for the purpose)
- To view the output you only need to specify the measure
The flow you specify should have an input called "well". Your flow should extract the channel from here and process it. The flow should output a ParticleMeasure to the output "pm"
Viewing the layout and images
Endrov tries to detect your input image layout. If it finds groups of channels called a1, a2, a3 etc it will assume it is a multi-well plate and show as such. Otherwise it will display a linear list of items. If the layout can be improved for your data, let us know how. The layout looks like this if you haven't specified anything to be shown:
If you you specify images then it looks like this:
At the bottom are controls for changing the Z (slice in stack) and frame (time). There are also controls for adjusting contrast and brightness (does not affect the data).
TODO thumbnail size
Viewing the extracted data
For each identified "particle" you will have a series of attributes, like intensity and volume. These attributes are chosen as Primary and secondary attribute in the drop boxes. Whenever only one attribute is used for viewing, it will be the primary attribute. For example, you can show a histogram of the primary attribute:
and you can show a scatterplot using also the secondary attribute on the y-axis:
You can also show "aggregations" of parameters such as the mean value, standard deviation, and so on. This is shown color coded from black (lowest value) to white (largest value). This is for example the mean intensity for particles in each well:
The mean value is probably the most useful parameter for quality control. Below is one example where dirt in a well causes the mean intensity to go off the chart:
Storage and post-analysis
Endrov is not a statistics program and we do not intend it to be one. Use the right tool for the job. For large-scale statistics, this means you should be using R. RStudio is a productive and easy to use graphical user-interface.
In the menus you find the option to export the output data as CSV or into an SQL database. You can however also store the output data in the OST file-format so that you do not have to re-calculate it the next time. It is likely that we will add more visualization capabilities in the future and you might find it convenient to be able to re-visualize your data right away.