DEDS

DEDS integrates multiple statistical metrics via a weighted distance framework to improve detection of differentially expressed genes in microarray data.

Key Features:

• Comprehensive statistical methods: Functions compute t statistics, fold change, F statistics, SAM (Significance Analysis of Microarrays), moderated t-test, and moderated F-statistics.
• Differential Expression via Distance Summary (DEDS): Synthesizes multiple statistical metrics using a weighted distance method to select differentially expressed genes.
• Comparative method support: Includes comparisons with t-test, eBayes (empirical Bayes t-test), TREAT (t-tests relative to a threshold), and AAA (assumption adequacy averaging).
• Microarray data focus: Operates on microarray expression data for differential expression analysis.

Scientific Applications:

• Differential expression detection: Identifies genes exhibiting differential expression across conditions or treatments in microarray experiments.
• Disease mechanism investigation: Supports analyses aimed at understanding disease-associated transcriptional changes.
• Biomarker discovery and therapy development: Aids discovery of candidate biomarkers and evaluation of targets for targeted therapy development.

Methodology:

Empirical evaluation on microarray data compares t-test, SAM, eBayes (empirical Bayes t-test), TREAT, and AAA across sample sizes; consistency is measured by percentage of overlapping genes (POG) and related POGR scores; power is assessed via simulations that replicate the multivariate distribution of microarray data; results report that moderated tests (SAM, eBayes, TREAT) perform better at small sample sizes with TREAT showing highest consistency, while for larger samples AAA shows higher power for small effect sizes and TREAT maintains lower power but better consistency.