Over the past few decades, accumulating evidence has highlighted CIN as a driving force behind the initiation and progression of various tumor types. The relationship between CIN and tumors is complex, involving both causative and adaptive elements. Since CIN is a fundamental feature of cancer understanding his significance is essential for elucidating the underlying mechanisms of cancer and exploring potential therapeutic strategies. The idea of my work starts from the article "A Pan-Cancer Compendium of Chromosomal Instability" , which has laid a critical foundation for understanding CIN, primarily through the analysis of copy number data. This thesis work takes a novel approach by harnessing the power of machine learning to predict the CIN signature identified in the article, employing methylation and gene expression data as alternative data sources to enhance our comprehension of CIN in cancer. The central aim is to determine whether CIN signatures, as previously elucidated using copy number data, can be replicated and validated using alternative molecular features and the research begins with the compilation of a rich dataset comprising methylation and gene expression profiles from diverse cancer cohorts, mirroring the pan-cancer approach. All the work was done using the R programming language.
Negli ultimi decenni, l'accumulo di prove ha evidenziato che l’instabilità cromosomica(CIN) è una forza trainante per l'inizio e la progressione di vari tipi di tumore. La relazione tra CIN e tumori è complessa, coinvolgendo sia elementi causali che adattativi. Poiché la CIN è una caratteristica fondamentale del cancro, la comprensione del suo significato è essenziale per chiarire i meccanismi sottostanti del cancro ed esplorare potenziali strategie terapeutiche. L'idea del mio lavoro ha origine dall'articolo "A Pan-Cancer Compendium of Chromosomal Instability", il quale ha gettato una fondamentale base per la comprensione dell'Instabilità Cromosomica (CIN), principalmente attraverso l'analisi dei dati di copy number. Questo lavoro di tesi adotta un approccio innovativo sfruttando il potere del machine learning per prevedere le signature di CIN identificate nell'articolo, utilizzando però dati di metilazione ed espressione genica come fonti dati alternative per migliorare la nostra comprensione di CIN nel contesto dei tumori. L'obiettivo principale è determinare se le signature di CIN, precedentemente chiarite attraverso i dati di copy number, possano essere replicate e validate utilizzando caratteristiche molecolari alternative. La ricerca inizia con la creazione di un ricco dataset comprendente profili di metilazione ed espressione genica da diverse coorti di pazienti oncologici, seguendo l'approccio pan-cancer. Tutto il lavoro è stato svolto utilizzando il linguaggio di programmazione R.
Approcci di machine learning per predire profili di instabilità genomica in campioni tumorali
ROTA NEGRONI, MARCO
2022/2023
Abstract
Over the past few decades, accumulating evidence has highlighted CIN as a driving force behind the initiation and progression of various tumor types. The relationship between CIN and tumors is complex, involving both causative and adaptive elements. Since CIN is a fundamental feature of cancer understanding his significance is essential for elucidating the underlying mechanisms of cancer and exploring potential therapeutic strategies. The idea of my work starts from the article "A Pan-Cancer Compendium of Chromosomal Instability" , which has laid a critical foundation for understanding CIN, primarily through the analysis of copy number data. This thesis work takes a novel approach by harnessing the power of machine learning to predict the CIN signature identified in the article, employing methylation and gene expression data as alternative data sources to enhance our comprehension of CIN in cancer. The central aim is to determine whether CIN signatures, as previously elucidated using copy number data, can be replicated and validated using alternative molecular features and the research begins with the compilation of a rich dataset comprising methylation and gene expression profiles from diverse cancer cohorts, mirroring the pan-cancer approach. All the work was done using the R programming language.File | Dimensione | Formato | |
---|---|---|---|
Marco_Rota Negroni.pdf
accesso aperto
Dimensione
14.71 MB
Formato
Adobe PDF
|
14.71 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/60034