Enhancing cancer diagnostics unveiling oncogenic gene fusions with oxford nanopore technologies long-read sequencing Karleena Rybacki

Format:

Book

Thesis/Dissertation

Author/Creator:

Rybacki, Karleena, author.

Contributor:

University of Pennsylvania. Bioengineering., degree granting institution.

Language:

English

Subjects (All):

Bioengineering.

Bioinformatics.

Genetics.

0202.

0715.

0369.

Local Subjects:

Bioengineering.

Bioinformatics.

Genetics.

0202.

0715.

0369.

Genre:

Academic theses

Physical Description:

1 online resource (231 pages)

Contained In:

Dissertations Abstracts International 87-12B

Place of Publication:

Ann Arbor : ProQuest Dissertations and Theses, 2026

Language Note:

English

Summary:

Gene fusions (GFs) are structural alterations that play a central role in cancer biology and precision oncology. Recurrent oncogenic GFs serve as diagnostic and prognostic biomarkers and direct therapeutic targets across diverse tumor types, making GF detection a routine component of clinical diagnostics. Currently, GFs are detected using short-read RNA sequencing-based fusion panels that target predefined sets of genes recurrent in fusions. While clinically validated and widely adopted, these panels are constrained by targeted gene boundaries, read-length limitations, and dependence on downstream computational reconstruction, limiting their ability to detect novel, rare, or structurally complex GFs. This also leaves an important diagnostic gap in tumors that test negative for GFs by targeted approaches. Furthermore, even when GFs are detected, the prioritization and interpretation of fusion candidates remain labor-intensive, inconsistent, and dependent on expert manual review. In this dissertation, I develop an integrated computational framework to improve both the detection and interpretation of GFs in cancer. First, I develop and evaluate a novel ensemble ONT-based GF detection pipeline, benchmarking its performance against an established short-read clinical fusion panel on samples with known GF status, demonstrating compatibility with existing short-read sequencing and improved turnaround times. Second, I extend this framework to clinically challenging fusion panel-negative cases by applying unbiased ONT long-read whole-transcriptome RNA sequencing, moving beyond the boundaries of targeted gene panels to enable untargeted GF discovery across the full transcriptome. Select candidate novel GFs identified from this analysis are orthogonally validated using in-vivo functional assays in Drosophila melanogaster to assess their oncogenic potential. Finally, I introduce FusionRank, a learning-to-rank (LTR) framework that integrates sequence-based, functional, and pathway-relevant features to prioritize GFs by their predicted pathogenicity, thereby reducing the burden of manual interpretation. Collectively, this work advances GF detection workflows by combining targeted and unbiased RNA-seq approaches, extends their application to clinically challenging cases, and introduces a LTR computational prioritization strategy that enhances the clinical interpretability of GF candidates in precision oncology

Notes:

Source: Dissertations Abstracts International, Volume: 87-12, Section: B.

Advisors: Wang, Kai; Ko, Jina Committee members: Shaffer, Sydney M.; Li, Marilyn M.

Ph.D. University of Pennsylvania 2026

Vendor supplied data

Local Notes:

School code: 0175

ISBN:

9798247982548

Access Restriction:

Restricted for use by site license