Expression Analysis Introduces Innovative New Comparison Tool.

DURHAM, N.C. -- Expression Analysis, a full-service genetics testing and services company specializing in the Affymetrix GeneChip(R) Brand technology, has developed a new two-group comparison analysis for differential expression. This tool allows researchers to reliably detect and estimate changes in expression between two experimental groups of three or more samples.

The comparison is useful for experimental designs that involve a baseline group and an experimental group such as control vs. treated, negative vs. positive, mutant vs. wild-type, or any two conditions from a multi-condition experiment.

By incorporating permutation analysis for differential expression (PADE(TM)), Expression Analysis is able to estimate a reasonable False Discovery Rate (FDR) for a set of potential differentially expressed transcripts. Assessing a set of transcripts is often a useful method for comparisons of this type since false positives frequently overwhelm true positives in microarray analysis where simple t-tests are performed on each transcript. According to Dr. Wendell Jones, Expression Analysis' Senior Statistician and developer of PADE, this is simply due to fundamental statistical issues with interpreting p-values when one is performing multiple parallel testing of tens of thousands of transcripts. "When R.A. Fisher developed t-tests in 1925, he quite obviously did not have present-day microarrays in mind."

"Permutation testing is based on "rerandomization" of the samples in order to build a reference distribution of the test statistic of interest. After a series of rerandomizations, typical values for the ordered test statistic are computed and then compared with the actual ordered test statistic from the experiment. With PADE, Expression Analysis uses a modified, robust t-statistic" says Jones.

"We are always investigating methods to improve the results we provide to our clients and are proud to offer this added flexibility to them at no additional cost," said Expression Analysis Founder and COO Steve Casey. "Our clients can choose either a cumulative or marginal FDR, or a combination of FDR with fold change, as thresholds to select transcripts of interest. This helps our clients move their projects to the critical next step."

Expression Analysis, Inc. (http://www.expressionanalysis.com) is a full-service microarray genetics testing and analysis organization that provides high-quality genomic processing and data analysis services. EA is one of the few U.S.-based authorized service providers for Affymetrix, Inc.