AlertGS: determining alerts for gene sets

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Gene expression studies aim to identify significant genes within specific gene sets.
• Modeling concentration/time-response relationships for individual genes is common.
• Current methods lack a global significance measure for entire gene sets.

Purpose of the Study:

• To extend the concept of gene-wise alerts to gene sets for global significance.
• To determine the earliest point of gene set enrichment in concentration or time-response data.
• To develop a robust methodology for analyzing gene set behavior in expression studies.

Main Methods:

• The AlertGS methodology utilizes a Kolmogorov-Smirnoff type test statistic.
• It transfers the concept of alerts from single genes to gene sets.
• The approach models concentration/time-response relationships for gene sets.

Main Results:

• Simulations demonstrate successful identification of a majority of true gene sets, particularly with fewer signals.
• False positive rates are manageable through decorrelation techniques.
• Gene set alerts are generally not overestimated, tending towards underestimation.

Conclusions:

• AlertGS offers a powerful tool for identifying enriched gene sets in gene expression data.
• The method provides a global significance measure and identifies the onset of enrichment.
• The AlertGS methodology is implemented and available for reproducible research.