The Dashboard

Once you see the Dash App running in your browser, you can inspect your data and run several tests and functions. If you don´t know how to achieve that have a look into the Running Dash Interface section.

Below you will find an image of the App with some descriptions underneath.

1 The Distribution Panel

The distribution panel shows how the protein is spread along the gradient. An averaging Kernel with specified size is used to calculate the amount of protein in each fraction. The Kernel size can be set via the Settings (4). You can also find a pseudo-Westernblot from the mean of your Replictas. If you want to inspect each replicate seperately you can trigger the Replicate Mode with the switch at the top of the panel.

2 The Data Table

As mentioned in the Data Prepatation step. You can include lots of additional data in the intensities file. Per Default only the first column from that file is shown on App startup. However, you can select additional columns via the Select Table Columns Input Field at the bottom. Once you calculated values using the functionality from (4), those values will also show up in the table. You can also click a row or use the arrow keys to navigate through your proteins and select which Protein to show in the Distribution Panel.

3 Sample Heatmap

This heatmap shows the distances between samples using the distance method specified in the Settings (4). If the treatment had some effect, the between group distances is supposed to be higher than the within group distances.

4 Settings

The Settings box is used to analyze your data. The purpose of the provided functionality is to analyze multivariate data.

Distance Method

Here you can select a distance metric to calculate the distance between samples. We strongly recommend to use the Jensen-Shannon-Distance here.

Kernel Size

Per default an averaging kernel is run over the fractions. This will reduce the variance of samples in case that fractionation did not work perfectly.

Get Score

Calculates the default values that you can use to rank your table. Those values are

Value	Explanation
ANOSIM R	The ANOSIM R Value ranges between -1 and 1. Proteins that show a different distribution upon RNase treatment will have a higher R value (ideally 1). This value can be used instead of a p-Value if you have a low number of replicates (approximately < 5).
Mean Distance	Mean Jensen-Shannon-Distance between treated and untreated replicates. This is a measure of effect size of the treatment.
shift direction	The direction in which the protein shifts upon RNase treatment.
RNase False peak pos	The weighted average fraction of the control from which the largest shift in probability mass towards the RNase treated samples is detected.
RNase True peak pos	The weighted average fraction from the RNase from which the largest shift in probability mass towards the control samples is detected.

Rank Table

This will ad a rank to the table considering the current sorting. For instance if you have an insufficient number of samples for p-value calculation via PERMANOVA or ANOSIM you might consider ranking the table based on a sorting using ANOSIM R and the Mean Distance.

Run ANOSIM

Performs ANOSIM to calculate whether RNase treatment leads to a difference in the distribution of the Proteins. You can select the number of permutations via the Input Field next to the button. The Default is 999.

Note

This is the recommended way to calculate a p-Value if you have a sufficient number of samples.

Export TSV

This will export the Data Table with all the values calculated using the buttons above.

Select Color Scheme

If you dont like the default colors you can change them here. There are several preset color schemes.

5 Dimension Reduction

Here you can see a dimension reduction of all shifts. On the x-axis you will see the relative fraction shift. This is caclulated via subtracting the average control shift position from the RNase treated ones. On the y-axis you can see whether the protein has a broader (more negative) or sharper (more positive) distribution after treatment.

Note

Proteins selected in the table will be highlighted in this plot and you can hover over data points to show the corresponding distribution.

3D

This will display three dimensions in the plot if toggled. Instead of showing the mean distance via the marker size it will add another axis for this value.

Cluster Method

Warning

Note that the clustering is currently disabled in the GUI and you can only use it via the Python API.

The Method that is used for clustering.

Adjust Cluster Settings

Opens a modal where you can adjust some settings of the clustering algorithms.