Decoding splicing variants in high-throughput sequencing: a functional validation approach integrating deep learning...

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• High-throughput sequencing identifies intronic variants in disease genes.
• Predicting the splicing impact of these variants, especially atypical ones, remains difficult.
• Deep learning and motif-based tools have limitations in detecting novel regulatory mechanisms.

Purpose of the Study:

• To evaluate the splicing consequences of intronic variants in neurodegenerative diseases.
• To assess the performance of various splicing prediction tools, including deep learning and motif-based approaches.
• To investigate the role of splicing enhancers and regulatory proteins like SRSF2.

Main Methods:

• Selection of nine intronic variants with uncertain splicing impact from neurodegenerative disease patients.
• Application of multiple complementary splicing prediction tools.
• Functional RNA studies including transcript analysis in blood and minigene assays.

Main Results:

• Splicing abnormalities were detected in eight of nine tested variants.
• All tested prediction algorithms showed limitations, missing certain splicing events or generating false positives.
• A deep intronic variant activated a splicing enhancer, potentially involving SRSF2 and leading to pseudoexon inclusion.

Conclusions:

• Splicing prediction tools are complementary but not fully sufficient for pathogenicity assessment.
• Transcript-level confirmation is critical for accurate interpretation of intronic variants.
• Integrating diverse prediction methods and functional validation is essential for diagnostic pipelines.