SUMO Scatterplot viewer - The Graphical User Interface

Here you will find a short description of the Graphical User Interface.

Main menu

Toolbar

Context menus

Main menu

File menu

Open tab delimited text data files, or SUMO-Scattterplot files containing additonally group display settings.

Import foreign data formts (FCS format).

Save as loaded data including custom groups / dsiplay settings as SUMO-Scattterplot files.

Export the graph as vector graphics in
- Scalable Vector Graphics (SVG)
- Windows Meta File (WMF)

Vector graphs may be sized without loss of quality and freely modified with suited editors (e.g. InkScape).

For details about supported file formats see Data files entry.

Edit menu

Find

Search the list of loaded data points.

Either search in :

Names / annotations
Data values

Matching datapoints are indicated with pink markers.
The list of matching datapoints is shown in the Session log tab sheet.

A dialog opens up.
Specify a search term.
From the drop down list you may recall a recently used search term.

By default search is performed on data points' names (e.g. gene symbols).

Enter a search key (e.g. "VEGFA") to search exatly this item.
Use wild cards ("*") to find partially matching datapoints:

"vegf*" to find "vegfa" and "vegfr" ...'
"*gf" to find "ctgt" and "vegf" but not "vegfa" nor "vegfr"
"*gf*" to find all of the above

To search the numerical data values (x,y,"size","color" specify:

"\x≥nnnn.nn"	all objects where X-valiue is equal or larger specified threshold
"\x≤nnnn.nn"	all objects where X-valiue is equal or smaller specified threshold
"\y≥nnnn.nn"	all objects where Y-valiue is equal or larger specified threshold
"\y≤nnnn.nn"	all objects where Y-valiue is equal or smaller specified threshold
"\s≥nnnn.nn"	all objects where size-value is equal or larger specified threshold
"\s≤nnnn.nn"	all objects where size-value is equal or smaller specified threshold
"\c≥nnnn.nn"	all objects where color-value is equal or larger specified threshold
"\c≤nnnn.nn"	all objects where color-value is equal or smaller specified threshold
"\r≥nnnn.nn"	all objects where RATIO x/y≥specified threshold
"\r≤nnnn.nn"	all objects where RATIO x/y≤specified threshold
"\lr≥nnnn.nn"	all objects where RATIO log2(x/y)≥specified threshold
"\lr≤nnnn.nn"	all objects where RATIO log2(x/y)≤specified threshold

Threshold MUST be a number with international dot (".") as decimal divider.
Non numeric values are converted to ZERO.

To combine a new search with a set of already selected data points specify as first search key:

"\or" : to keep the existing selection and add new matching datapoints (logical OR)
"\and" : select the intersection of the old selection and the newly matched datapoints (logical AND)

E.g. select all growth factors data with x-value between 1000 and 2000:
- first search: "\x≥1000"
- next search: "\and x≤2000"
- next search: "\and *gf*"

Clear selection

Un-select all data points.

Alternatively, drag an empty selection with the mouse.

Copy data

Copy data from the Scatterplot as tab delimted text into clipboard.
All vales(x,y,X/Y-error,Color,Size,Grop-ID,Name) are copie4d.

Such a copied list may look like:

X	            Y	            Error-X	    Error-Y	    Size	    Color	   Group	Name
128546.48438	115674.15625	1820.84790	11942.00098	34.01107	39.77178	0	    1
 30870.58594	133569.12500	2894.34741	13599.37695	 3.44957	 9.55785	0	    2
-41971.60938	-54666.37500	3154.14136	 5431.87061	 6.53209	 3.38831	0	    3
-34792.36719	-68183.10938	3170.89355	 2756.95752	58.64328	63.78699	0	    4
-18017.51563	-98872.61719	1449.07239	 8313.46094	25.48250	19.61693	0	    5
 96628.62500	 42455.34375	 576.08539	 3467.49487	 2.25695	75.79591	0	    6
 ...

You may copy either:

All data contained within Scatterplot
Selection: only values from presently selected datapoints.

Paste data

Paste data from clipboard.

Paste single column data:

x-data
y-data
x-error
y-error
marker size
marker color
name / comment

Data points my be separated by TAB or NEWLINE characters.
Obviously, names/ comments shoud NOT contain internal TABs or NEWLINE characters.
Empty / non numeric data fields are ignored
Always paste x-data first !!

Take care, all pasted colums have identical size and structure / order. Otherwise your data will get scrambled up.
In case a individual data value (e.g. an error value is missing) supply a "0" for this data cell.

Paste multi column data:

x/y/name
x/y/x-error/name
x/y/y-error/name
x/y/x-error/y-error/name
x/y/x-error/y-error/size/color/name

Paste data as multi-line ASCII text

One datapoint / line
Data values separated by tabs.
Non-numeric / mising values are converted to ZERO
Additional data fields are assigned to NAME.
Empty lines or lines with invalid x/y values are ignored.

Data table

Show data contained in Scatterplot as data table spread sheet.

Refresh

Just redraw the graph.
Or press "F5" function key.

View menu

Here you find some options to cunstomize how data are displayed.
By default Scatterplot displays x/y data.
In case you have aditonal data values, you may use those to size symbol and define symbol color.
Additonally you may supply error values for x/y-axes.
Check form View menu:

Size
Size values should be ≥0
Scatterplot scales the values linear into the interval 0..Maxvalue
Color
You may supply:
- only positive values: values are linear scaled into the interval 0..Maxvalue
- signed values: ScatterPlot finds absolut MAX value. Data are linear scaled into the interval -MAX..0..Max
X-error
Supply positive values
Y-error
Supply positive values

to view these additional values (if supplied).

Examples:

Default display:	x/y + size:
x/y + size/color:	x/y + error:

Density map

In case you have a large number datapoints within you scatterplot, the dense overlapping symbols may hide overall data distribution.

In such a case it may be better to transform the scatterplot into a Density map and Hide symbols:

Symbols:

Density map:

For each (screen) pixel the number of overlaying datapoints is transformed into overlayed 2D-gaussian shaped "clouds".
Color/Intensity visualize data multiplicity/density.

In the example, 10000 random (gaussian distributed) data points are shown.
A White (no data) => blue (max density) color-palette is used.
A 31-pixel / 7-SDev gaussian kernel was used for transformation (default).
Goto Main menu | Preferences | Density map to adjust Gaussian kernel parameters or select other color palettes.

Selection groups, ... may still be shown in the density maps:

Symbol

Variable Max symbol size

Define the Maximum symbol size (in pixel, e.g. "50") for the symbols when a "size" variabe is used.
I.e. size data values are autoscaled:

Min size value => 3 pixel
Max size value => defined maximum symbol size (e.g. 50)

Variable color palette
Select the palette to display "color" values

Fixed monochrom, colors:

White => Red
White => Green
White => Blue
White => Black

Fixed two colors:

Green => Red
White => Green
Blue => Red

Custom 2-color palette:
A two-color selection dialog opens up:

Click the Low / High panels and select any of the preset colors or define a new one.

Fixed three colors:

Green => White => Red
Blue => White => Green

Custom 3-color palette:
A three-color selection dialog opens up:
Click the Low / Middle / High panels and select any of the preset colors or define a new one.
Three color palettes may be most useful for signed "color" values.

Examples:

Monochrome:

2-Color:

3-color:

Diagons

Show / hide a Main diagon in the graph

Show / hides 2x/4x/8x-regulaton Side diagons in log/log scaled scatterplots.

Zero lines

Show / hide zero lines for x-/y-axes in the graph.

Names of Selected / Groups

Check to show names / comments of those datapoint just rigth to the datapoint which are member of custom groups or the present selection:

Connection lines

Connect individual datapoints with linear line segments.
I.e. make your ScatterPlot look like a line chart.

Requires, that x-values are sorted.
To sort da data on x-values select Main menu | Utilities | Sort data by x-values.

Direct line:

Right step:

Left step:

With unsorted data (x-values) the graph may look confusing:

View selection list

Open a a spread sheet windows to view the list of datapoints matched by your lst searc a painted ROI.

Update all other viewers

Assume you analyzed several replicated samples after same treatment.
Now you may show scattreplotss for each replca agianst a common control.

It might be interesting to see whether up-regulated datapoints (selected in one scatterplot), show up in the other scatterplots clustered together in the same region.

To test this, check Main menue | View | Update all other viewers.

Select some datapoints in one scatterplot (search or draw a ROI).
The respective datapoints are selected in all other scatterplot too:

Selected source:

Replica 2:

Replica 3:

The example indicates:
The "upregulated" datapoints from scatterplot-1 are randomly spread within the two replicas.
Thus, a suspected directed regulation from sample-1 can not be confirmnd with the additional two replicates.

To perform such an analysis, it is required, that order and number of datapoints in the multiple scatterplots are identical.

Utilities menu

Swap axes

Exchange data between any pair of data colums.

From the dialog select the two columns you want to exchange.

Sort data by x-axis

Sort all data points based on x-values in increasing order.

Demo data

Generate a set of random data.

Different underlying data distributions may be selected:

Linear:

Gaussian:

Half circle:

Waves:

A parameter dialog opens up.

Define a few parameters for the dataset:

Number of datapoints
Y-offset
Linear y-trend (slope)
noise level

Window menu

A few options to handle Scatterplot windows.

Arrange

If you opened muliple Scatterplots, SUMO can try to arrange them systematically:

Tile: Size / position all Scatterplots so they cover the whole desktop.
Cascade: Position all Scatterplots with small increasing x-/y-offset.

Close all

Close all presently open Scatterplots.

New scatterplot

Open a new Scatterplot copying all data.

New empty Scatterplot

Open a new empty Scatterplot.

Transform menu

Preferences menu

Scaling

Define how to scale x/y values

Log-X / Log-Y

Display x/y vlaues log10 transformed.
Only possible if ALL data values are ≥ 0

Symmetric

Use same common min/max for x-/y-axis.

Min/Max

Define min/max-valies for x-/y-axis manually.

Selection

Define how to draw a Region Of Interest (ROI) with mouse keeping left button pressed:

Rectangle: rectangle with bounds parallel to x/y-axis.
Rhombus: rectangle with bounds paralles to main diagon.
Freehand: Free irregular shape.

Quadrant analysis:

Split the Scatterplot into four neighboring quadrant, with the mouse position as central connection point (i.e. click mouse at desired position.

Number of datapoints / quadrant are sown (absolute and relative) as well as "Cross table" enrichement p-valaue (Fischer exact test):

Mean Standard deviations for all data columns are shown in the Log-Tabsheet.

Join selection groups

Obvious.

Customize selection groups

Same as toolbar button GP.

A dialog opens up:

Select a group and customize its display preferences.

Save group settings

Save presently defined display settings for all groups.
Thus you may set-up e.g. various project specific group setting and easily load them lateron.

Load group settings

Reload previously saved group settings.

Density map

Set peferences for density map.

Kernel parameters

Define how the 2D-gaussian distributed "clouds" are computed.

Size: Size of the kernel (default=31 screen pixel).
As wider as more homogenious the map may look.
Adjust SDev accordingly.
SDev: width of the gaussian shape (Default=5).
To get a smooth map the the Gaussian profile must fit into the kernel.
Linear / Log: Log transformation may improve the visualizatioon when high density-"peaks" ther are in the dataset.

Just try some combinations:

Size=11,Sdev=2,Linear:
Size=21,Sdev=4,Linear:	Nice and smooth for a small picture. But not if you expand the image
Size=91,Sdev=10,Linear:	Click to see large view.
Size=21,Sdev=7,Linear:	Gaussan profile to wide for computation kernel. Click to see large view.