Two-way AIC: detection of differentially expressed genes from large scale microarray meta-dataset

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Detecting differentially expressed genes (DEGs) from DNA microarray data is crucial in biomedical research.
• Conventional DEG detection methods often analyze single datasets, limiting the identification of experiment-specific genes.
• A meta-dataset approach is needed to simultaneously compare gene expression across multiple experimental conditions.

Purpose of the Study:

• To propose a novel method, two-way Akaike Information Criteria (AIC), for simultaneous detection of significant genes and experiments within a meta-dataset.
• To enhance the identification of experiment-specific DEGs by analyzing gene expression across diverse conditions.
• To evaluate the utility of the two-way AIC method for gene interaction estimation.

Main Methods:

• Constructed a "meta-dataset" by aggregating gene expression data from public databases.
• Developed and applied the "two-way AIC" (Akaike Information Criteria) method for simultaneous gene and experiment significance detection.
• Utilized Pseudomonas aeruginosa operon genes as test data for evaluating method specificity.

Main Results:

• The two-way AIC method demonstrated high specificity in detecting experiment condition-specific DEGs, specifically operon genes.
• Compared to traditional methods like t-rank/F-test, RankProducts, and SAM, two-way AIC exhibited superior specificity.
• The case study on Pseudomonas aeruginosa confirmed the method's effectiveness for identifying specific gene sets.

Conclusions:

• The two-way AIC method achieves high specificity for detecting operon genes within microarray meta-datasets.
• This novel approach facilitates the identification of experiment-specific DEGs across multiple conditions.
• The two-way AIC method shows potential for estimating mutual gene interactions.