Tutorials

4.1 Clustering with SOTA

If you have your own data you can go directly to the server: Sotarray Server

1st Part:

Go to data sets page clicking here:

DNA-arrays: Public Data Sets

Click on Diauxic shift:

Diauxic Shift Data Set

(Optional)  Note in the different files the steps in the treatment of the data

Choose a file (data_norm.txt, for example) and send it to the server by clicking here:

sota/cluster

Select SOTA server:

Send To Sotarray

Alternatively you can send the original data set to the preocessor by clicking on:

preprocess

You will have to log-transform the data, to remove flat patterns and maybe to standardize the patterns.

[Tutorial of the Preprocessor]

Select Unrestricted Grow option

Click on Run.

Send result to TreeView server to see the tree:
 

Send To TreeView

Accept default parameters and press Run

Go back by clicking here:

Change Parameters

Uncheck Gene Names and change to 1 the value of Vertical Separation

Press Run again.

---

2nd Part:

Close the window and go back to the Sotarray result screen

Press Change Parameters button to modify`parameters:
 

Change Parameters

Select Variablity Threshold (%) option and set the value to 90 (default).

Press Run again.

Send the results both to TreeView and al SotaTree:
 

Send To SotaTree

Send To TreeView

Compare the results

4.2 Classical hierarchical clustering (UPGMA)

If you have your own data you can go directly to the server: Cluster Server
In data sets page choose one of the experiments. DNA-arrays: Public Data Sets
Choose a file and send it to the server by clicking on: sota/cluster Send it to the Cluster server

Send To Cluster

Alternatively you can send the original data set to the preocessor by clicking on:

preprocess [Tutorial of the Preprocessor]
Accept default values (UPGMA and correlation) and press Run
Send result to TreeView to display the tree:

Send To TreeView

Press the change parameters button, uncheck Gene Names and set the Vertical Separation to 1.
Press Run

4.3 Self-Organizing Maps

If you have your own data you can go directly to the server: SOM Server
In data sets page choose one of the experiments. DNA-arrays: Public Data Sets
Choose a file and send it to the server by clicking on: som/somtree Send it to the SOM server:

Send To SOM

Alternatively you can send the original data set to the preocessor by clicking on:

preprocess [Tutorial of the Preprocessor]
Accept default values and press Run

4.4 SOM-Tree

If you have your own data you can go directly to the server: SOM-Tree Server
In data sets page choose one of the experiments. DNA-arrays: Public Data Sets
Choose a file and send it to the server by clicking on: som/somtree Send it to the SOM-Tree server:

Send To SOM-Tree

Alternatively you can send the original data set to the preocessor by clicking on:

preprocess [Tutorial of the Preprocessor]
Accept default values and press Run
Clusterize the result of SOM by clicking on:

Clusterize it

4.5 Play with more examples

Go to data sets page by clicking here: DNA-arrays: Public Data Sets As you have repeatedly been told, if you have your own data, feel free of using our server: Clustering Resources

4.6 Analysis of the results: data mining

Select a cluster with a few genes and look for their function in a database like SRS or Swiss-Prot. Try to assign a global function to the cluster. Try to do the same using FatiGO and compare the results.