Tutorial (for Windows Users)

Installation

  1. Install the latest release of GNU R (at least 2.8).
  2. Download the latest A-MADMAN all-in-one package
  3. Click on A-MADMAN-setup.exe and follow the instructions in particular:

Technical details (skip if you don't care)

The all-in-one package is quite big because for ease of deployment bundles a few software packages: All important settings are stored in localconf.py
For the all-in-one package we choose reasonable defaults. In particular we provide a SQlite db preconfigured with necessary tables and default administrative data preloaded. Feel free to change the defaults...

Concepts and terminology

The all-in-one package creates a user named admin with password admin (as the name suggest can also access the administrative interface). A default group named test is also created. The test groups owns the default project named also test. (admin is obviously part of group test)

Data Retrieval

Data to retrieve must be specified in a configuration file. The syntax is pure python. You define a simple data structure that specifies the names of series and samples to download.
In the installation directory there is an example configuraton file named georc.example.
	
data={'GSE1004': {'samples': ['GSM15807', 'GSM15822', 'GSM15823', 'GSM15824', 'GSM15825', 'GSM15826', 'GSM15827', 'GSM15828', 'GSM15829', 'GSM15830', ]}, 'GSE1786': {'samples': ['GSM30842', 'GSM30843', 'GSM30844', 'GSM30836', 'GSM30837', 'GSM30838', ]}, }
You can specify the names of samples if you need only a subset of samples from a series or 'all' if you need all for example:
	
data={'GSE1004': {'samples': 'all'}, 'GSE1786': {'samples': 'all'} }

Launch the A-MADMAN console from the Start Menu and start the download process with:

python manage.py geoget --georc georc.example
The download process will take a while... (depending on how many series you selected and how fast is your network).
If something goes wrong reissue the command and the process will restart from where it left.

(Meta) Data import

In the A-MADMAN console run this command:
python manage.py geotodb --georc georc.example --project test
This command will create series and samples objects in the database and will import the associated metadata (in the test project).

Login and first look

From the Start Menu launch start-amadman and point your favourite browser to http://localhost:8000/amadman/.
In the main menu on the top of the page click projects.
A login form will appear. Login with username admin and password admin.
Then from the projects drop down menu select the test project (the only one).
Now the complete menu is available:

Annotation of samples

Tag creation

We'll create three tags to annotate our samples.
Click Tags in the main menu and add three tags using the form.
Name them: young, old and sedentary (a brief description is mandatory).

Adding tags to samples

Now you can use the tags to annotate the samples.

Start with series GSE1004: Now go on with series GSE1786:

Baskets creation

Click Extract on the top menu.
The interface for query creation is shown.
Click on young and the press the preview button.
In a new tab/window you can preview the results of this simple query (all samples tagged with young)
Now close the tab/window and press go
Now give a name (name it 'youngs' ) and description to the newly created basket and press the save button.

Click again Extract and compose the following query:
'old and not sedentary'
Preview the query results. Then create and save a new basket and name it 'well_trained'

Individuals assigment

In this tutorial we will proceed with a rapid automatic assignment: a new individual is assigned to each sample. In a real meta-analysis you would have to check carefully if more samples refer to the same individual and annotate properly this situation. See here for detailed instructions.
Click Assign in the top menu choose a series clicking on the number of samples left to assign and then press the auto button. Do the same for the other series.

Vanilla Analysis

Click on Baskets on the top menu.
Mark the two baskets you just created youngs and well-trained and press the create analysis button.
Name the analysis 'test1', leave the default and only workflow vanilla and press the save button.

The analysis is now queued for execution and must be executed by another a-madman component the job server.
Select start-job-server from Start Menu -> A-MADMAN
The job server will continue to run until the window is closed, waiting for analyses to execute.

The analysis status will change from waiting, to running to done.
When the status is done you can click on the analysis name and see the execution log, download the R workspace and see the generated R code.

Defining a custom workflow

We'll define a workflow to substitute rma signal reconstruction with gcrma.

Click Custom WorkFlows on the top menu and press the new button.

Give the new workflow the name gcrma, fill in a brief description and substitute the default text in the template field with the following code and press the save button
	
{% extends "basic.rtmpl" %} {% load R %} {% block cdf_flavour %} flavour="affy" {% endblock %} {% block signal_reconstruction %} load_or_get_from_bioc("gcrma") {% for chip_name in chip_names %} load_or_get_from_bioc("{{chip_name}}probe") ai <- compute.affinities("{{chip_name}}") eset.{{chip_name}} <- gcrma(batch.{{chip_name}},affinity.info=ai) {% endfor %} {% endblock %} {% block additionalcode %} load_or_get_from_bioc("preprocessCore") ieset=metanorm(ieset) {% endblock %}

In this case we changed only the signal reconstruction entry point and leaved the rest of the workflow unaltered. See the general documentation for details.

Running an analysis with a custom workflow

Click on Baskets on the top menu.
Select the two baskets and press the create analysis button.
Give a name to the newly created analysis, select gcrma from the Workflow drop menu and press the save button.