On Bioinformatic Methods for the Detection of Alternatively Spliced Variants for Clinical Diagnostics Dina Issakova

Format:

Book

Thesis/Dissertation

Author/Creator:

Issakova, Dina, author.

Contributor:

University of Pennsylvania. Biology., degree granting institution.

Language:

English

Subjects (All):

0306.

0541.

0715.

0769.

0984.

Local Subjects:

0306.

0541.

0715.

0769.

0984.

Physical Description:

1 electronic resource (115 pages)

Contained In:

Dissertations Abstracts International 87-07B

Place of Publication:

Ann Arbor : ProQuest Dissertations and Theses, 2025

Language Note:

English

Summary:

Exome sequencing (ES) is the standard of care for patients with suspected Mendelian disorders,yet its diagnostic yield remains limited at 31%. A major contributor to this gap is the inability of conventional ES analysis to capture RNA splicing defects-a tightly regulated process in which pre-mRNA segments are joined into distinct isoforms. Splicing abnormalities are implicated in an estimated 15-50% of pathogenic variants, including synonymous substitutions. Robust tools to detect splicing aberrations from RNA-seq are therefore essential for improving diagnostic outcomes.For clinical adoption, such tools must accurately detect diverse splicing events under varying levels of inclusion, handle confounding factors, remain consistent with known biology, and be computationally feasible. Pioneering methods such as LeafCutterMD and FRASER have advanced this space but face limitations in accuracy, efficiency, and usability. To address these challenges, the Barash lab developed MAJIQ-CLIN, a pipeline for identifying patient-specific splicing aberrations relative to large control cohorts. However, the accuracy, practicality, and clinical utility of these approaches have not been systematically evaluated-an essential step for their integration into diagnostic workflows. This dissertation aims to fill that gap, advancing RNA-seq-based splicing detection toward clinical application. Chapter 1 reviews the biology of alternative splicing, its role in disease, and current diagnostic algorithms: FRASER, LeafCutterMD, and MAJIQ-CLIN. Chapter 2 benchmarks these tools across aberration types and variant inclusion levels, and compares their outputs to predictions from SpliceAI. Chapter 3 examines clinical accessibility, comparing computational resource usage, control cohort design, and confounder correction strategies, and provides guidelines for parameter selection in diagnostic practice. Chapter 4 applies MAJIQ-CLIN to Undiagnosed Diseases Network (UDN) data, independently confirming a diagnosis previously identified by optical genome mapping and uncovering a putative novel transcription start site in MCPH1, suggesting a potential gain-of-function variant. This chapter also presents a differential splicing analysis of TBCK syndrome, highlighting the utility of splicing detection for mechanistic insight into rare neurodevelopmental disease. Finally, Chapter 5 discusses the limitations of this work and outlines directions for future research.By integrating reproducible benchmarking, compatibility with widely used datasets, and novel biological findings, this dissertation delivers a practical framework for RNA-seq-based splicing diagnostics. Ultimately, it aims to support the broader adoption of splicing analysis in clinical genomics and to contribute to improved diagnosis and treatment for patients with rare disease and their families

Notes:

Advisors: Barash, Yoseph; Bhoj, Elizabeth Committee members: Murray, John; Gao, Ziyue; Wang, Kai; Gregory, Brian

Source: Dissertations Abstracts International, Volume: 87-07, Section: B.

Ph.D. University of Pennsylvania 2025

Vendor supplied data

Local Notes:

School code: 0175

ISBN:

9798276004914

Access Restriction:

Restricted for use by site license