Mining high dimensional transcriptomic data to unravel the causes and consequences of genomic instability in cancers

    Research output: ThesisThesis fully internal (DIV)

    112 Downloads (Pure)

    Abstract

    Genomic instability is an enabling hallmark of cancer that comprises the progressive accumulation of copy number alterations (CNAs). We need to understand how the CNAs facilitate the genomically unstable cancer types to acquire the cancer hallmarks. In chapter 3, by reanalysing >34,000 gene expression profiles, we revealed the degree of transcriptional adaptation to CNAs in a genome-wide fashion. We observed that ~10%, ~50%, and ~40% of genes have a low, moderate, and a high degree of transcriptional adaptation to CNAs, respectively. Also, it is essential to safeguard the process of DNA replication to maintain genomic instability. In chapter 2, we identified a six-gene signature of oncogene-induced replication stress (NAT10, DDX27, ZNF48, C8ORF33, MOCS3, and MPP6). These signature genes identified DLBCL, ovarian cancer, TNBC and colorectal carcinoma as cancer subtypes with high levels of oncogene-induced replication stress. In chapter 4, consensus-independent component analysis (c-ICA) on expression profiles of 1,089 epithelial ovarian cancer samples identified a cohort with the worst survival (15% of the patients), defined by high activity of the transcriptional footprint with similarities to neuronal development. Incomplete functional information about these identified genes from chapters 2-4 can limit the interpretation of the results. Predicting gene functions can overcome this limitation. In chapter 5, we show that usage of ICA-derived transcriptional components 1) provides more confident gene functionality predictions, 2) improves predictions when new members are added to the gene-sets, and 3) is less affected by gene multifunctionality in comparison with principal component analysis derived transcriptional components.
    Original languageEnglish
    QualificationDoctor of Philosophy
    Awarding Institution
    • University of Groningen
    Supervisors/Advisors
    • van Vugt, Marcel, Supervisor
    • Fehrmann, Rudolf, Co-supervisor
    Award date11-May-2022
    Place of Publication[Groningen]
    Publisher
    Print ISBNs978-94-6419-500-2
    DOIs
    Publication statusPublished - 2022

    Cite this