Simultaneous detection and estimation of trait associations with genomic phenotypes

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• Genomic phenotypes (DNA methylation, chromatin accessibility) reveal cellular transcriptional and regulatory activity.
• Dense measurement technologies capture spatial structures where nearby genomic sites exhibit similar phenotypes.
• Comparing these phenotypes across conditions (e.g., cell types, diseases) is crucial for biological insights.

Purpose of the Study:

• To propose a novel statistical method, Joint Adaptive Differential Estimation (JADE), for comparing genomic phenotypes.
• To leverage the inherent spatial structure of dense genomic data for improved analysis.
• To simultaneously estimate group-average profiles and detect differential regions between experimental conditions.

Main Methods:

• Joint Adaptive Differential Estimation (JADE) method development.
• Leveraging spatial correlation in genomic phenotype data.
• Simultaneous estimation of smooth group average profiles and differential region detection.

Main Results:

• JADE effectively estimates underlying genomic phenotype profiles.
• JADE accurately detects regions with differential average profiles between groups.
• Performance validated through biologically plausible simulation settings.

Conclusions:

• JADE offers a robust approach for comparing genomic phenotypes across diverse biological contexts.
• The method's ability to leverage spatial structure enhances differential analysis.
• Application to skeletal muscle cell differentiation demonstrates practical utility in identifying differential methylation.