Annotating pathogenic non-coding variants in genic regions.

Area of Science:

• Genomics
• Medical Genetics
• Bioinformatics

Background:

• Accurate genetic variant interpretation is crucial for understanding disease.
• Current methods often miss pathogenic non-coding variants, specifically synonymous and intronic types, hindering disease risk assessment.

Purpose of the Study:

• To introduce the Transcript-inferred Pathogenicity (TraP) score for reliably identifying pathogenic non-coding genetic variants.
• To develop a tool that overcomes limitations of existing methods in detecting disease-associated synonymous and intronic variants.

Main Methods:

• Development of the TraP score utilizing sequence context alterations to predict variant pathogenicity.
• Validation of TraP using public datasets of synonymous and intronic variants, assessing performance via AUC.
• Application of TraP to exome data from epilepsy family trios to identify disease-associated variants.

Main Results:

• TraP accurately distinguishes pathogenic from benign synonymous (AUC=0.88) and intronic (AUC=0.83) variants with high specificity.
• The score identifies extremely rare variants with lower minor allele frequencies compared to missense variants.
• Analysis of epilepsy trios revealed pathogenic synonymous variants in known epilepsy genes using TraP.

Conclusions:

• The TraP score is a valuable tool for identifying pathogenic non-coding variants, improving gene discovery and personal genome interpretation.
• TraP outperforms existing methods in detecting pathogenic non-coding variants, offering a significant advancement in genetic diagnostics.
• The availability of a web server with pre-computed scores facilitates broader application of TraP.