A game changer for bipolar disorder diagnosis using RNA editing-based biomarkers

Area of Science:

• Neuroscience
• Genetics
• Bioinformatics

Background:

• Differentiating Bipolar Disorder (BD) from unipolar depression is clinically challenging due to overlapping depressive symptoms.
• Misdiagnosis leads to delayed treatment and poorer patient outcomes.

Purpose of the Study:

• To discover RNA editing biomarkers for distinguishing BD from unipolar depression.
• To develop a diagnostic tool using machine learning for early BD identification.

Main Methods:

• A-to-I RNA editome analysis in a discovery cohort to identify differentially edited variants.
• Candidate biomarker validation in a larger cohort using stringent criteria and pathway analysis.
• Machine learning models applied for discriminating depression subtypes.

Main Results:

• Identified 646 differentially edited variants in 366 genes in depressed patients versus controls.
• A panel of 8 gene biomarkers distinguished depressed patients from controls with high accuracy (AUC 0.930).
• A combination of 6 biomarkers differentiated BD from unipolar depression with high sensitivity (90.9%) and specificity (84.6%).

Conclusions:

• RNA editing variants are associated with depression subtypes.
• An AI-powered tool utilizing RNA editing biomarkers can accurately differentiate bipolar disorder.
• This diagnostic approach promises to reduce misdiagnosis delays and facilitate timely treatment for bipolar disorder.