A practical framework for predicting splicing single nucleotide variants in exome sequencing.

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Splicing variants are key contributors to Mendelian disorders.
• Predicting the pathogenicity of splicing variants remains a significant challenge in genetic diagnostics.

Purpose of the Study:

• To develop a simplified framework for evaluating pathogenic splicing single nucleotide variants (SNVs).
• To align with the 2023 ACMG/AMP guidelines and ClinGen recommendations for variant classification.

Main Methods:

• Developed a scoring system assigning priority scores (-10 to 14) to SNVs in open reading frame regions.
• Validated the framework using pathogenic splicing SNVs from the Human Gene Mutation Database and common SNVs from gnomAD.
• Compared the framework's discriminatory power against SpliceAI.

Main Results:

• The framework demonstrated superior discrimination compared to SpliceAI alone (AUC 0.991 vs. 0.983, P = 2.11 × 10⁻²³).
• Identified pathogenic splicing variants in known genes (COL2A1, PDHA1, MECP2, JAKMIP1) in 1257 patients with unresolved diagnoses.
• Suggested potential candidate disease-causing genes (UBN1, NFE2L1).

Conclusions:

• The developed framework simplifies the pathogenicity assessment of splicing SNVs.
• This method enhances the detection of splicing variants through exome sequencing, aiding in diagnosing genetic disorders.