Detection of selection signals on regulatory units across human cancers

Tumor progression is dominated by two evolutionary forces: first mutagenesis, which provides the heritable variability where, secondly, natural selection acts. The main challenge of cancer genomics is to identify the somatic mutations that drive the tumorigenesis, the drivers, from the vast majority...

Descripción completa

Detalles Bibliográficos
Autor: Rodríguez Galindo, Miguel
Tipo de recurso: tesis doctoral
Estado:Versión publicada
Fecha de publicación:2023
País:España
Institución:CBUC, CESCA
Repositorio:TDR. Tesis Doctorales en Red
OAI Identifier:oai:www.tdx.cat:10803/689480
Acceso en línea:http://hdl.handle.net/10803/689480
Access Level:acceso abierto
Palabra clave:Selection inference
Somatic evolution
Mutational processes
Cancer
Tumor
Non-coding selection
Somatic variants
Càncer
Genòmica
Inferencia de selecció
Evolució somàtica
Processos mutacionals
Tumors
575
Descripción
Sumario:Tumor progression is dominated by two evolutionary forces: first mutagenesis, which provides the heritable variability where, secondly, natural selection acts. The main challenge of cancer genomics is to identify the somatic mutations that drive the tumorigenesis, the drivers, from the vast majority of neutral variation, the passengers. A decade of careful surveying of tumor DNA has revealed a multitude of protein-coding drivers, several of which have been used as therapeutic targets. However, many tumors do not exhibit any of these known exonic driver events, leaving a gap in our knowledge. Recently, large efforts have been made querying the remaining non-coding part of the genome. Intriguingly, the role of non-coding somatic mutations still remains largely less well understood than its protein-coding counterpart. These regions are specially challenging: poorly annotated, dominated by abnormal mutational processes that act as confounders, and with a broader mutational target than coding regions. To address these specific challenges, I developed an approach that identifies selection in regulatory regions. Regulatory units are tested as a whole to increase the statistical power, and the impact of mutations is evaluated through biophysical models for transcription factor binding sites. Finally, a mutational model accounts for mutation rate variability at several scales. With this approach I find known and new putative cancer drivers on a harmonized cohort of 7586 whole cancer genomes and conclude that non-coding selection is ubiquitous in cancer evolution.