The objective of this tool is to provide network based interpretation for compound list.
The KEGG REACTION database is used as refference knowledge.
Here is how it works:
- You need a list of compound Ids. Ids are KEGG compound identifiers.
- To submit a new job, go to the "Home" page.
- Upload a list of compunds and submit your job by pressing the "submit" button.
This is what the server will do next:
- When the job is started, the server will first check the list of supplied Ids.
The results would be reported in the table which will summirize:
the number of read Ids,
the number of recognized KEGG Ids.
- For recognized KEGG Ids TICL will provide network model.
In simple terms TICL will try to infer maximal spanning "Semi noninterrupted"
subnetwork that covers the supplied list of compounds. Semi noninterrupted means that
we fix in advance the number of missing compounds between any two compounds from supplied list
while we still consider them to be connected. This extensive definition
of noninterrupted network allow to account for false negative compounds
that are missing for some technical reasons in the list.
Therefore, TICL will provide several network models
in respect to parameter that defines the maximal number of missing compounds.
-
TICL will provide an estimate of statistical significance for the inferred model.
The p-value provided (estimated by Monte Carlo simulation) by TICL indicates
a probability to infer the same size network model from a random list of compounds
of the same size as analyzed input list.
The Output of the server:
- After the submission of the data you will be redirected to the main results page.
The computation may be time consuming and to get the results you need to update this page.
We recommend you to bookmark this page and update it in approximately 5-15 minutes.
-
The main results page will contain links to the results.
First the report of input data is provided.
Second, a table "Enriched subnetworks"
which reports the statistics of the inferred network models
in respect to the parameter "Maximum distance between compounds".
Each row of this table correspond to parameter
which defines the maximal number of missing compounds between any two compounds
from the input list to consider them connected.
The column "Number of nodes" reports the number of compounds in the inferred network
from the corresponding compound list.
The column "p-value" reports the estimated statistical significance
of the inferred model: the probability to get the same size network model
for a random compound list.
The first column provides links to visualize the corresponding network model.
Attention: to be able to use this facility you must to have Java to be installed on your computer
and allow your browser to run Java applets.
-
After clicking on visualization link you will be redirected to visualization page.
In your window you should see network visualization.
In most cases the metabolic network is big for visualizing them on line.
The best way is to download and save on your computer the text file provided
(follow the link "Download model for interactive visualization with Medusa").
Simply open this file in Medusa program and use all Medusa visualization capacities
to produce graphical figures.
This server is bound to evolve in the near future to aquire additional functionalities.
All comments are welcome!
PS: As this is a freely accessible non-for-profit web service,
it comes of course with absolutely no warranty (see imprint).
This is a brief syllabus describing the different job options and input file formats.
- Input list:
-
The input format is one compound Id per line
Here we will post frequently asked questions about the TICL web server
How long does a TICL job take?
A typical job takes about 5-15 minutes.
Most of this time is spent to compute background distribution
and estimate the p-value of the inffered models.
The best way is to bookmark the main result page and update it in 5-20 minutes.
My job still does not yield the expected results!
We are constantly trying to improve our server.
If you encounter any unexpected problems, please do not
hesitate to contact Alexey Antonov for help.
|
