[BioC] ArrayQuest at MUSC

[Gary Argraves]

Greetings to bioconductor users,

The purpose of this letter is to introduce you to ArrayQuest. ArrayQuest™ is
a web-based program for the analysis of DNA microarray data. ArrayQuest™ is
designed to apply various types of analysis scripts (i.e., Bioconductor
statistical and graphical methods) to microarray data stored in the MUSC DNA
Microarray Database and the Gene Expression Omnibus (GEO).

We have developed a software suite that includes four modules that work in
conjunction to mediate the archiving and analysis of DNA microarray data.
The four modules of this suite are as follows:

   1. µArrayDB™ module: µArrayDB™ is a program for archiving DNA microarray
data derived from experimentation performed through our DNA Microarray
Facility. The data for µArrayDB™ is stored on our Internet front-end server,
a Dell PoweEdge 440. µArrayDB™ and ArrayQuest™ (see next) both utilize MySQL
server to store meta data such as project information (scanned and analysis
project information), cRNA target information and process controlling
information.

   2. ArrayQuest™ analysis module: this module allows a client to select
microarray information from the µArrayDB™ database for analysis. The
ArrayQuest™ analysis paradigm can be thought of as an analysis process
controller in which a user specifies an analysis process that includes a
selectable analysis method (i.e., a particular R script such as RMA) and
related input for that method (i.e., Affymetrix .CEL files). ArrayQuest™ has
been designed to automatically distribute loads across a computing cluster.
Before assigning a clients job, ArrayQuest™ will query a load control
program that will indicate which CPU in the cluster is best suited to
analyzing the job at the time. ArrayQuest presently uses Linux operating
systems RedHat and Fedora Core2. ArrayQuest™ is operated through web
browsing portals such as (i.e., Internet Explorer, Safari, Mozilla). The R
package that is used to run many ArrayQuest methods can be compiled to
utilize the increased performance that a 64 bit architecture offers.

   3. Analysis job-distribution module: ArrayQuest™ communicates with a
job-distribution module to send an analysis job to a CPU within the backend
analysis cluster, that is ready to accomplish the job. A key application
used in the analysis system is a statistics package entitled R. Most
analysis jobs that are launched by clients will take a considerable amount
of time for the results to be returned. ArrayQuest™ will notify the client
by email as well through flagging status within the process database, the
current state of any process being executed.

   4. Upload module: This module will allow off site clients to import data
into µArrayDB™ or to a private database. This will allow the µArrayDB™
database to grow. When the database grows then our client base will grow and
in turn our analysis requirements will grow. For this reason we need the CPU
Analysis Custers. µArrayDB™ is designed to protect user data and not allow
it to be released into the public domain without authorization, as data
security is a major issue in the field. The µArrayDB™ module has been
describe in a published article (Argraves et al., 2003).


URL:
http://proteogenomics.musc.edu/quickSite/arrayQuest.php?page=home&act=manage

Please direct your interest in this system to:

W. Scott Argraves, Ph.D., Professor
Medical University of South Carolina
Dept of Cell Biology and Anatomy
173 Ashley Avenue
Charleston, SC 29425
Email: argraves at musc.edu

or to

garyArgraves at sbcglobal.net

Sincerely,
Gary Argraves