Background: We previously introduced a method based on post hoc pairwise comparisons to analyze gene expression responses. This method utilized directed graphs to represent gene response to all treatment pairs. It has been found useful in identifying structure-activity relationships among drugs and differentiating genes sharing similar functional pathways. Directed graphs are descriptive, visually expressive and can benefit subsequent functional analysis. Results: mDAG is a web-based software package based on this established method for the analysis, visualization, and interpretation of patterns of responses in gene expression data involving multiple treatments. Genes with the same directed graph patterns hypothetically share similar biological function, which may be further analyzed using external tools. To facilitate subsequent functional analysis, several well-known tools have been incorporated into mDAG to help users explore hypotheses about gene function and regulation. This tool is useful for any studies that analyze comparatively response patterns in gene expression data with multiple treatments (chemicals, cell types, etc.). Availability: The (server/personal/demo) software is freely available at
http://cetus.cs. memphis.edu/mdag.