Arrow plot: a new graphical tool for selecting up and down regulated genes and genes differentially expressed on...

Area of Science:

• Bioinformatics
• Genomics
• Statistical Genetics

Background:

• Microarray data analysis commonly identifies differentially expressed genes using distance measures.
• Biological sample heterogeneity, including molecular subtypes, can lead to missed gene expression differences with standard methods.
• Bimodal or multimodal distributions in gene expression often indicate underlying sample mixtures.

Purpose of the Study:

• To propose a novel graphical tool for identifying genes with differential expression, including those in subgroups.
• To address limitations of standard statistical methods in detecting complex gene expression patterns.
• To offer a more comprehensive analysis of gene expression profiles.

Main Methods:

• Developed a new graphical tool based on the overlapping coefficient (OVL) and area under the receiver operating characteristic (ROC) curve.
• Implemented the methodology in the open-source R software.
• Applied the tool to publicly available and simulated datasets.

Main Results:

• The proposed method identified differentially expressed genes with bimodal or multimodal distributions missed by standard procedures (Welch t-statistic, fold change, rank products, etc.).
• Compared favorably against other methods like area between ROC curve and rising area (ABCR) and test for not proper ROC curves (TNRC).
• Demonstrated the ability to detect genes with differential expression in subgroups and consider different variances between samples.

Conclusions:

• The arrow plot is a novel, flexible, and useful tool for analyzing gene expression profiles from microarrays.
• The method provides a more comprehensive analysis by detecting multimodal distributions and handling varying sample variances.
• Graphical analysis of different types of differentially expressed genes is an advantage.