Automated Microarray Pipeline (AMP)

AMP is a collection of data analysis software that will normalize and statistically analyze raw microarray data. AMP includes algorithms derived from the TM4 suite of tools developed at TIGR and DFCI and now available at http://tm4.org. AMP is a pipeline, meaning that analysis methods are linked together in a chain from start to finish, and no user intervention is required during the processing of data. Users submit raw un-normalized data via a web interface to AMP. Raw microarray data files are uploaded to the AMP server, where they are stored and processed. The user selects the methods used for data processing. Resulting data are viewable in the AMP web pages, but also are downloadable.

Login

You know what to do!

Register

User ID and password are created when you register. In order to manage your data, AMP needs to know who you are. Fill in the requested information and click submit (fields marked with * are required). AMP collects project leader (PI) information in addition to user information [OK, what’s our justification for this??]. Once your information has been submitted, you will be sent an email that you can use to activate your user ID.

Note: both your user name and email address must be unique on our AMP site.

Java WebStart

In order to use the AMP web pages, you must have Java WebStart installed on your computer. Information on Java WebStart can be found here. These same instructions can be accessed by clicking the information (i) button on the AMP ‘Upload Data’ web page.

Upload Data

The first step in processing data with AMP is to create a ‘study’. Do this by typing a name in the ‘New Study’ field. The Dataset name field is also required to help distinguish analyses that you start on the same day. Indicate (by selecting a value from the dropdown list) whether your data should remain private, or can be made available for query through our WebArray search page (not currently available). (Data are maintained in a secure, private database; data are made available only when users indicate that it can be made public.) Then click ‘Upload Data’. This will start WebStart, and open the Upload panel.

If you have already created an AMP study, then there are 2 options:

  1. select the study name from the ‘Existing Study’ list and select upload data, or
  2. click on the study name in the ‘uploaded study list’ area.

Note: AMP currently accepts only Affymetrix CEL files. We are planning to provide analysis methods for 2-color microarray data in a future version.

Upload Panel

The Upload panel allows a user to send data files (eg CEL files) to AMP. To upload data files follow these steps:

  1. To upload Affymetrix click on the ‘Browse’ button of this panel. This will open an ‘Open’ dialog box which will allow you to browse to the files you wish to upload.
  2. Once you click ‘Open’, the list of selected files will be displayed. AMP detects the chip type from the CEL files; different chip types are highlighted with different colors. In order to upload a batch of files, they must all be of the same chip type. You can remove files from the list by selecting them and clicking ‘Remove Selected’. You can clear the list entirely by clicking ‘Remove All’.
  3. Once all the selected files are of the same chip type, the Upload button will be activated. Click on Upload to begin sending your data files. You can cancel the upload by clicking ‘Stop’, if desired.

Data Normalization and Analysis

On this page you define the processing that will be done for your data.

  1. Select a normalization method. (Note: Only RMA normalization is available at the current time.)
  2. Select a scaling method. If you wish to scale your results by mean or median values, select those options and provide a value on which the data will be centered (typically 0).
  3. Select a statistical analysis method. Each of the statistical methods has additional options that can be selected; these options are described in the MeV manual. Select the necessary options, and then click 'OK'.
  4. To cluster the array data based on GO annotation, select EASE and provide the requested information in the EASE dialog box.

If you do not wish to use a particular method, simply select 'none'. Each method has a notes field that you can use to identify your data analysis results. Once done selecting options, click the 'Submit' button at the bottom of the page.

Data Analysis Results

To review results of the analysis of your data, click on the link to the specific results on the right side of the screen. The result sets are listed with dates, times and analysis IDs to help you select the data to review. To save the results on your computer, right-click the desired dataset link and select “save link as”. The data are formatted as tab-delimited text which can be opened with MS Excel.

The number of files generated by the analysis will differ between analysis types. Some of the files generated are listed below:

RMA.txt

results from RMA normalization

xxx-Sig-Genes.txt

genes/probesets with significant p-value

xxx-Nonsig-Genes.txt

genes/probesets with non-significant p-value

xxx-POP.txt

the population file for the analysis

xxx.log

log file for the analysis

To further analyze a set of data, click the radio button to the left of the test that you requested, and then click the 'Submit Data for Data Analysis' button. This will open the Data Normalization and Analysis page (described above) for you to submit another analysis.

The downloaded data can be easily imported into MeV (where many more analysis methods are available). MeV can be obtained at: tm4.org. To load the RMA.txt file into MeV, use: File > Load Data. On the Expression File Loader window, change the drop-down menu at the top of the page (labeled "Load expression files of type") to "Tab Delimited, Multiple Samples Files". Then browse to select the downloaded file. Once selected, click 'Load'.