Output: Transcript Co-Response Analysis (Summary)

All co-response analyses will present the data in tabular format. You will be presented with the correlation coefficient a rank (according to correlation coefficient) a gene description and some statistical values like p-value, number of pairs, confidence interval, and power of test (you can show and hide these parameters by clicking 'SHOW' or 'HIDE'). Moreover you will be presented with the mutual information, converted into a distance range d(MI), which is a non-linear measurement of the association / relation of two genes. You can also color the result set by clicking or decolorize by clicking . In this case p-value and power of test will be colored according to the following scheme. Best values are depicted in red, intermediate ones in yellow and borderline cases in green. Worst values are not colored (see color definition).
If you click on the correlation coefficients (these are links) you will be presented with a new page where you can verify and display the pre-calculated co-response using various methods. Moreover you will be presented with links to external annotation databases.
At the end of each co-response table you will also be presented with two pie-charts showing the distribution of the hits into classes. One the one hand (top pie chart) you will be presented with a distribution according to keywords identified in the gene description and on the other hand (botton pie char) you will be presented with a normalized distribution of genes according to functional classes (functional neighbourhood search).

Output: Transcript Co-Response Analysis (Tutorial)

If you have submitted your gene(s) of interest and your gene(s) is (are) in the selected / default matrix you get the results in tabular HTML format.

Here is a 'standard' output labelled with comments. The output may slightly differs among the different query types.


You will be presented with the rank (according to correlation coefficient), the second gene, the resulting correlation coefficient, a gene description and some statistical values like number of pairs (the pairwise measurements for gene a and b)[n], p-value [pvalue], confidence interval [CI], power of test [power], the mutual information as distance measure [d(MI)], and the normalized euclidean distance [d(E)]. The additional parameters (CI, power, d(MI), and d(E)] can be obtained if you click on SHOW or HIDE.

Typical output after request. Additional statistical parameters are not visible.

You can show and hide these additional statistical parameters by clicking 'SHOW' or 'HIDE' at the help area on top of the table.

You can also color the result set by clicking or decolorize by clicking . In this case p-value and power of test will be colored according to the following scheme. Best values are depicted in red, intermediate ones in yellow and borderline cases in green. Worst values are not colored (see color definition).

The result set table will be ordered by descending levels of co-regulation. Note however that you can click on the table headers of every column to resort the table according to that column.

At the end of each co-response table you will also be presented with two pie-charts showing the distribution of the hits into classes. One the one hand (top pie chart) you will be presented with a distribution according to keywords identified in the gene description and on the other hand (botton pie char) you will be presented with a normalized distribution of genes according to functional classes (functional neighbourhood search).

Output: Color Definition

The values are colored as follows:

Output: Functional Neighbourhood Search

At the end of each co-response table you will also be presented with two pie-charts showing the distribution of the hits into classes. One the one hand (top pie chart) you will be presented with a distribution according to keywords identified in the gene description and on the other hand (botton pie char) you will be presented with a normalized distribution of genes according to functional classes (functional neighbourhood search).
For explanation both pictures are adjoin: left word searching, right functional neighbourhood search.

(E.g. for Arabidopsis 35 classes, thus each piece of the pie would represent approximately 3% if no enrichment would take place) In the figure (above) the classes 4,5,7,10 seem to be enriched to a normal background, possible indicating that genes from these classes are the major interaction partners of the queried gene.

Word Search:
This approach based on identified keywords in the gene description for the output. If a keyword identified in the gene description the counter for this keyword will be counted up with 1. There is no normalization to the number of the same keyword in the description of all represented genes in the corresponding data matrix.

Functional Neighbourhood Search:
This approach based on the assignment of the genes onto functional classes. Each gene is assigned onto one or more functional classes (categories), in the worst case 'unknown'. The count up value depends on the number of assignments, e.g. a gene with 2 assignments will get a count up value of 0.5. If this gene in the obtained output a 0.5 is added to class 1 as well as to class 2. For each class we obtained the sum of this count up values for the retrieved output. To get information about the enrichment in one or more classes the obtained sum for one category is divided by sum for this category, computed over all genes represented in the data matrix.