web analytics

A Computational And Therapeutic Platform For Identifying And Correcting Unproductive Splicing To Restore Gene Expression

Published:
Lead Inventor: Xiaochang Zhang
Print or Save to PDF

SUMMARY

Advanced computational tools to identify faulty RNA splicing events and design of antisense oligonucleotides to restore normal protein production in genetic diseases, especially those affecting the brain

The Unmet Need: Analytical pipelines to uncover previously undescribed unproductive splicing events and new therapeutic targets

  • Alternative splicing is a fundamental process in eukaryotic gene expression, enabling a single gene to produce multiple transcript isoforms and thereby increasing proteomic diversity. However, this process is tightly regulated, and disruptions can have significant biological consequences. One critical regulatory mechanism is nonsense-mediated mRNA decay (NMD), which eliminates transcripts containing premature stop codons, often arising from alternative splicing events. In the context of haploinsufficient diseases—where a single functional gene copy is insufficient for normal function—unproductive alternative splicing events that trigger NMD can further reduce protein output from the wild-type allele, exacerbating disease phenotypes. As a result, there is a pressing need for technologies that can accurately identify and therapeutically target these unproductive splicing events to restore functional protein expression, particularly for genes implicated in neurological and developmental disorders.

  • Current approaches to identifying and correcting unproductive alternative splicing events are limited by several factors. Traditional computational methods for predicting NMD events often rely on simplistic rules, such as the "50-nucleotide rule," which do not capture the complexity of transcript isoform diversity or the nuanced features influencing NMD efficiency. These methods frequently yield false positives or negatives, missing many developmentally regulated or tissue-specific splicing events that are relevant to disease. Moreover, therapeutic strategies to modulate splicing, such as antisense oligonucleotides, have typically focused on well-characterized exons or mutations, leaving a vast landscape of unannotated, disease-relevant AS-NMD events unaddressed. The lack of robust, high-throughput tools to systematically identify and prioritize these events has hindered the development of targeted therapies for a wide range of genetic disorders.

The Proposed Solution: Computational platform employing machine learning to predict and classify unproductive alternative splicing events from transcriptome data and design ASOs for haploinsufficient genes

  • The faculty inventor developed a comprehensive platform for identifying and therapeutically redirecting unproductive alternative splicing events that trigger nonsense-mediated mRNA decay (AS-NMD), with the goal of restoring functional protein expression in haploinsufficient genes. At its core is the computational pipeline, which analyzes transcriptome sequencing data—both long- and short-read—to predict and classify AS-NMD events. The platform simulates transcript isoforms, calculates critical distances between stop codons and exon-exon junctions, and employs a machine learning model trained on transcript features to score NMD efficiency. The platform has mapped thousands of developmentally regulated AS-NMD events in mouse and human brains, many linked to neurological disorders.

  • For therapeutic intervention, it enables the rational design of antisense oligonucleotides (ASOs) that suppress the inclusion or exclusion of unproductive exons, as demonstrated by increased functional GRIA2 and FLNA transcript levels in cell-based assays. The platform’s ability to systematically catalog and validate thousands of AS-NMD events empowers researchers to target a broad array of previously inaccessible splicing events. Its demonstrated success in designing and validating ASOs for specific disease-relevant genes, along with detailed protocols and sequence data, positions it as a transformative tool for developing novel therapeutics aimed at restoring wild-type protein expression in genetic disorders.

ADVANTAGES

ADVANTAGES

  • Sophisticated computational pipeline integrating sequencing data and machine learning for accurate prediction and classification of AS-NMD events, outperforming traditional methods

  • Identifies thousands of developmentally regulated AS-NMD events, including those linked to neurological disorders, enabling targeted therapeutic interventions

  • Enables design of specific antisense oligonucleotides (ASOs) that effectively suppress unproductive exon inclusion or exclusion, increasing functional transcript levels in disease-relevant genes

  • Provides a comprehensive catalog and detailed methodology facilitating replication, optimization, and broad application in treating haploinsufficient and genetic diseases

APPLICATIONS

  • Gene therapy
  • Drug development
  • Drug discovery
  • Diagnostics
  • Validated across multiple experimental models and cell types, demonstrating robust and generalizable therapeutic potential