I. System requirements
II. Tilescope capacity
III. Using Tilescope
IV. Important usage notes

Tilescope (tl-skp) is a fully integrated data processing pipeline for analyzing tiling array data. In a completely automated fashion, it will normalize signals between channels and across arrays, combine replicate experiments, score each array element, and identify genomic features.

I. System requirements

To run Tilescope all you need is a Java-enabled web browser.

The Jave virtual machine (JVM) used by the browser should be of version 5.0 or higher. The system has been tested and demonstrated to work successfully with Firefox and Internet Explorer on Windows XP, Firefox and Safari on Mac OS X, and Firefox on Linux.

OS Browser
Windows XP Firefox, IE
Mac OS X Firefox, Safari
Linux Firefox
- On Windows XP, to enable JVM 5.0 for either Firefox or Internet Explorer, you need
1. Remove any old Java installations from you Windows system through "Start > Control Penal > Add or Remove Programs".
2. Remove any residual files and file folders in C:\Program Files\Java.
3. Download J2SE 5.0 from Sun and install it with default settings.

- On Mac OS X, to enable JVM 5.0 for Safari, you need
1. Download J2SE 5.0 from Appleand install it.
2. Open "Applications > Utilities > Java > J2SE 5.0 > Java Preferences" and select "J2SE 5.0" from the drop-down menu labelled "Use version" under the "Applets" heading.

- On Linux, to enable JVM 5.0 as a plugin for Firefox, yor need
1. Download J2SE 5.0 from Sunand install it.
2. Link installed JVM to Firefox as a plugin. In the Firefox plugin directory (for example, /opt/firefox-1.5/plugin/), create a symbolic link like this:
ln -s /jdk1.5/jre/plugin/i386/ns7/
/jdk1.5/ should be replaced by the path to the J2SE 5.0 installation on your Linux system.

You may need to relaunch your browser to use the newly-installed JVM. Depending on what web browser you are using, you may need to clear its cache to force a download of the latest version of the Tilescope applet.

II. Tilescope capacity

Currently due to server CPU and memory limitations, at any monent Tilescope can actively process one (1) data set, accept three (3) data analysis requests by storing them in a waiting queue, and handle three (3) concurrent data file transfers.

III. Using Tilescope

Tilescope is easy to use.

If you are using Tilescope to process an experiment data set for the first time, you are most likely to follow these steps:

1. First you need to transfer your data files to the Tilescope server. To do so, click "Upload..." button, browse to and select the files in the file selection dialog window. If no error occurs, the file transfer starts immediately and its status is given in the "Pipeline message" text area. Do NOT interrupt an active file transfer process.

2. After file transfer is finished, enter slide IDs directly in the "Slide IDs" text field (ignore "Load parameters from file" panel), separate IDs with white spaces, and then click the "Enter" button.

3. In the "Dye swap" table, click the cell to select appropriate dye swap (yes or no) options for each slide. No dye swap is the default selection.

4. Select the genome that was tiled on the microarray.
Code Genome
HS16 Homo sapien, build hg16
HS17 Homo sapien, build hg17
CA Candida albicans
SB Saccharomyces bayanus
SC Saccharomyces cerevisiae
SM Saccharomyces mikatae
NA Not available

5. Select the normalization method.

6. Set the sliding window size, tile length; set pseudo-median signal threshold, and P value threshold for filtering each scored tiles.

7. Select feature identification method and set relevant parameters for the method you selected.

Code Method
mgmr Max gap min run
ipi Iterative peak identification
hmm Hidden Markov model
8. Review settings of all
parameters. If they are what you prefer, click Run button.

9. On average it can take up to 15 minutes to process a three-replicate data set. Wait for your job to finish. When it is done, a web page with the results will be shown in a new browser window.

IV. Important usage notes

1. If the array data file is available in both gff and pair formats, the file in gff format should be transferred since gff is a more standardized format and thus less problematic for processing.

2. For Affy tiling array data (cel + array design files), the users need to change the names of their cel files before file transfer. The test and control sample file should be renamed to id_532.cel and id_635.cel respectively. We suggest id to take the form "user's name initial-today's date-a serial number". For example, an id can be "ZDZ-20060728-1". Please note the pair of test and control files should have the same serial number.

3. File transfer process extracts only relevant data items from the original data files, and also converts the file format to gff on Tilescope server.

4. Tilescope is NOT a file server. File directories storing users' data files are cleaned regularly. The users are responsible for their original data files.

5. The right genome should be selected to get the correct up-stream and down-stream gene annotation of the hits identified by the 'mgmr' method.

6. Quantile normalization is efficient and has been demonstrated to outperform other normalization methods. Thus it is recommended for data normalization both between channels and across slides.

7. For human and yeast data sets, 1000 bp and 400 bp are sensible window sizes to be used respectively.

8. Increasing pseudo-median threshold or p-value threshold will disqualify more tiles and thus less hits will be identified.

9. You may need to disable the popup-blocking feature of your web browser for this web site to view the result web page. If somehow this cannot be done, you can still view the result web page by typing "" into your web brower address text field.

Powered by . Comments? Write to zhengdong.zhang at
2006-2011 © ZDZ Lab, Albert Einstein College of Medicine