A Markov random field-based approach for joint estimation of differentially expressed genes in mouse transcriptome...

Area of Science:

• Neuroscience
• Genomics
• Statistical Modeling

Background:

• Mouse brain RNA-Seq data offers insights into neurodevelopment.
• Analyzing gene expression dynamics across time and cortical layers is crucial.
• Small sample sizes in developmental studies limit traditional statistical power.

Purpose of the Study:

• To develop a statistical method for identifying differentially expressed genes in mouse brain development.
• To address the challenge of small sample sizes in RNA-Seq data analysis.
• To leverage similarities across cortical layers, time points, and sexes for enhanced detection.

Main Methods:

• A Markov random field (MRF)-based statistical approach was developed.
• Model parameters were estimated using an efficient Expectation-Maximization (EM) algorithm with mean field approximation.
• The method incorporates between-layer similarity, temporal dependency, and sex similarity.

Main Results:

• The proposed MRF model significantly improves the power to detect differentially expressed genes compared to marginal analysis.
• Simulations and real data analysis validated the model's effectiveness.
• The method identified biologically relevant findings in the mouse brain RNA-Seq dataset.

Conclusions:

• The novel MRF-based statistical methodology enhances the analysis of neurodevelopmental gene expression data.
• This approach is particularly valuable for studies with limited sample sizes.
• The method provides a powerful tool for uncovering dynamic changes in brain development.