Changing the Coloring Scheme

Color by Expression

This option colors genes according to their normalized expression values and trustworthiness. To color your genes by expression, select Colorbar > Color by Expression .

Expression

The vertical axis of the colorbar represents expression levels on a continuous scale. Using the default colors, red indicates overexpression, yellow indicates average expression, and blue indicates underexpression. Genes are colored by their expression level in the selected condition as indicated by the condition line. If you have specified the parameter on the horizontal axis to be continuous, expression levels in between conditions will be interpolated.

Trust

The horizontal axis of the colorbar indicates the degree to which you can trust your data, where dark or unsaturated colors represent low trust, and bright, saturated colors represent high trust.

You can assign trust values for each gene when you load your experiment or allow GeneSpring to create trust values automatically (the latter numbers are listed in the Gene Inspector, in the Control column). To enter your own numbers, see The Experiment Wizard. The following are the guidelines by which GeneSpring automatically creates trust values:

In two-color experiments, the trust value is usually the control channel (typically Cy5), unless you do a per-chip normalization in which case it is:

(the control channel) x (the median of the control channel) x
(the median of the signal channel)

For Affymetrix and other one-color experiments, the trust value is constructed based on the normalizations you have chosen. If you accept the default normalizations for Affymetrix data (use distribution of all genes using the 50th percentile and normalize to the median for each gene), then trust is:

(the median value of the chip) x (the median value of the gene)

If you choose to use distribution of all genes using the 50th percentile and normalize to sample(s), trust is calculated as follows:

(the median value of the chip) x (the average of the gene's measurement in control samples)

GeneSpring automatically interprets trust for Affymetrix data, specifying 500 as data that is most trustworthy, 150 as moderately trustworthy, and 50 as least trustworthy. For other data, you will need to enter these numbers manually. Consult the manuals for your array scanning software or hardware and estimate these trust levels based on the detection limit and noise levels for any given measurement.

To set the trust interpretation:

Right-click the colorbar.
  1. Click Set Range .
  2. Enter values for High Control Strength, Medium Control Strength, and Low Control Strength.
  3. Click OK .

Color by Significance

Data is colored based on how far the gene is over- or underexpressed (relative to a normalized expression level of 1), in terms of the standard error of the measurement. The standard colorbar is replaced with a colorbar ranging from +3σ to -3σ. The standard error model is based on the Global Error Model, if the Global Error Model is turned on. (For more information about the Global Error Model, see Global Error Models.) Otherwise the standard error is based on the standard deviation of the replicate data for a particular gene and condition (for information about the calculation of this error, see the Gene Inspector).

To color your genes by significance, select Colorbar > Color by Significance .

Color by Static Experiment

This option allows you to color your experiment by a single condition. The vertical axis of the colorbar represents relative intensity on a continuous scale. In the default coloration, red indicates overexpression, yellow indicates average expression, and blue indicates underexpression. The horizontal axis of the colorbar indicates the degree to which you can trust your data, where dark, or less intense, color represents low trust, and light, or more intense color, represents high trust (for information about trust, see Trust).

To Color by Static Experiment

Click the + sign to the left of your experiment in the navigator.
  1. Click the + sign to the left of your experiment interpretation.
  2. Right-click over the condition you wish to color by and select Set Static Experiment .

To deselect color by static experiment, go to the Colorbar menu and select a different coloring scheme.

Color by Venn Diagram

This option colors genes based on their membership in one or more gene lists in a Venn diagram. For information about creating Venn diagrams and using them for analysis, see Making Lists with the Venn Diagram.

Color by Parameter

This option colors genes based on the value of parameters. This coloring scheme is best suited for use with Graph view and Bar Graph view when different conditions are indicated with discrete symbols.

To Color by Parameter

Select Experiments > Change Experiment Interpretation .
  1. Choose the parameter(s) you wish to color by and click Color Code for that parameter. Click SAVE to create a new interpretation.
  2. Select Colorbar > Color by Parameter .
 
The NIH Spinal Cord Study colored by parameter

No Color

This option allows you to view genes with no coloration, showing all genes in gray.

To implement this option, select Colorbar > No Color .

Color by Classification

This coloring scheme allows you to color-code the genes by some previously defined knowledge about them. You can use a folder of lists to color by classification or a classification method such as k-means or SOM.

Coloring a Previously Saved Classification

You can use a previously saved classification for coloring.

Open the Classifications folder by clicking its icon.
  1. Select a classification by right-clicking over the name.
  2. Select Set as coloring scheme from the pop-up menu and GeneSpring will automatically update to reflect the new coloring scheme.

The colorbar will show the names of the sets present in the chosen classification.

 
A Split Window, colored by Classification

Split Window and Color by Classification

You can also use the Split Window feature with the Color by Classification scheme.

Select a gene list to view.
  1. Right-click over a folder or a previously saved classification and select Use as Classification .
  2. Right-click over that folder again and select Split Window > Both .

Color by Secondary Experiment

The Graph and Scatter Plot displays lend themselves to being colored in many different ways because the display presents expression levels of the genes through the entire experiment. These are the only views in which you may choose to color the genes by a secondary experiment. This means the color of each gene line graphed correlates to the expression level of that gene in a different experiment, at the point in the second experiment marked by the secondary scroll bar.

From the navigator, open the Experiments folder by clicking on its icon.
  1. Position your cursor over an experiment (not the one currently displayed) you would like to use for coloration.
  2. Right-click and select Set Secondary Experiment from the pop-up menu.

The coloring scheme of the genome browser will be shown in the colorbar on the right. There will be two versions of the animation controls in the Experiment Specification Area.

Changing the Experimental Data Range

Before you change the experimental data range, you will need to select Colorbar > Color by Expression .

Right-click over the colorbar and select Set Range from the pop-up menu.
  1. Reset the values determining the intensity of the colors used by the genome browser.
  2. Click OK .

There are six categories you can change:

High Expression

High expression refers to the normalized expression of your genes, it is the vertical axis of the color bar. The default for this is 6.0.

Normal Expression

For most normalization procedures in GeneSpring the data will be normalized to 1.0. The default for this is 1.0.

Low Expression

For most normalization procedures in GeneSpring the data will not have negative numbers. The default for this is 0.0.

High Control

High control refers to the control strength of your genes. It is represented by the horizontal axis of the colorbar. The default for this is 200.0.

Medium Control

The default for this is 100.0.

Low Control

The default for this is 50.0.

For example, you could change the usual range of an experiment to high = 10, normal = 5 and low = -2 resulting in a very different color scheme once you click OK .

There is no Edit > Undo ( Ctrl+Z ) function for this type of change. If you want to return to your previous coloration scheme, you must re-open the Experiment Data Range pop-up window and type in your old values.

For more details on trust, please see Trust. For more details on normalization, please see Normalizing Options.

Changing the Default Colors

You can change the colors used by GeneSpring to display the genes. This will not affect the interpretation of your data, although it might help you to make genes more visible on-screen or make it easier to print screen shots.

Select Edit > Preferences .
  1. In the drop-down menu, select Colors .
  2. Select the type of information whose color you would like to change and click the Change button.
  3. Adjust the slider until the color you want is displayed in the preview window at the top of the Preferences window.
  4. Click OK .

For more details about the other options in the Preferences window, please refer to The Preferences Window.