Topological Sorting of Glioblastoma Patient Genes Based on Mutation Variations Using Quantum and Classical Computing Systems
Loading...
Authors
Chakrabarty, Mousumi
Issue Date
2026-05
Type
Dissertation
Language
en
Keywords
Healthcare Innovation & Delivery , deep learning , glioblastoma patient genes
Alternative Title
Abstract
Glioblastoma multiforme is a highly aggressive brain cancer characterized by complex muta-tion patterns that challenge traditional computational analysis. Identifying critical genetic driv-ers and their hierarchical relationships is difficult due to the high dimensionality and interde-pendencies within genomic data, limiting advances in precision oncology. The purpose of this study was to design and evaluate a hybrid quantum–classical deep learning framework for topological sorting of glioblastoma patient genes based on mutation variations. The study used publicly available somatic mutation data obtained from the cBioPortal for Can-cer Genomics, which provides open-access cancer genomic datasets derived from The Cancer Genome Atlas. These datasets are de-identified and distributed under open-access data sharing policies that permit use for academic research. The methodology followed a structured data sci-ence pipeline including preprocessing, feature engineering, and interaction modeling using mu-tation co-occurrence relationships. A Quadratic Unconstrained Binary Optimization model with a cardinality constraint was developed and solved using exact enumeration, a genetic algorithm, and a quantum approximate optimization algorithm. Results showed that the optimized gene subsets differed from prevalence-based selections, indi-cating interaction-based optimization. The genetic algorithm consistently achieved optimal so-lutions, while the quantum approach produced competitive but variable results. The findings demonstrate that hybrid optimization provides a scalable approach for genomic data analysis.