Future directions for high-throughput splicing assays in precision medicine

Area of Science:

• Clinical Genetics
• Molecular Biology
• Bioinformatics

Background:

• Classifying variants of unknown significance (VUS) poses a significant challenge in clinical genetics.
• Approximately one-third of disease-causing mutations are linked to pre-mRNA splicing, necessitating accurate splicing mutation classification.
• Large-scale patient sequencing studies generate novel variants faster than they can be classified, creating a bottleneck in variant interpretation.

Purpose of the Study:

• To compare high-throughput splicing assays for mitigating the variant classification bottleneck.
• To analyze a dataset of approximately 5,000 variants using the Massively Parallel Splicing Assay (MaPSy).
• To evaluate machine learning models for predicting splicing effects of variants, including the winning MMSplice model.

Main Methods:

• Utilized the Massively Parallel Splicing Assay (MaPSy) to analyze ~5,000 variants.
• Participated in the Critical Assessment of Genome Interpretation (CAGI) challenge, where machine learning models were submitted for splicing effect prediction.
• Compared the performance of MMSplice against existing software for variant classification.

Main Results:

• The Massively Parallel Splicing Assay (MaPSy) was employed to analyze a substantial dataset of variants.
• The MMSplice machine learning model, developed for the CAGI challenge, demonstrated superior performance in predicting splicing effects compared to existing software.
• Identified limitations of current high-throughput splicing assays and proposed methods for improvement.

Conclusions:

• High-throughput splicing assays and advanced machine learning models like MMSplice are essential for efficient variant interpretation in clinical genetics.
• Further development of these assays is needed to address challenges such as tissue-specific splicing and intronic variant classification.
• Enhanced methods for variant interpretation will significantly benefit the field of clinical genetics by improving diagnostic capabilities.