Inferring the tumour’s evolutionary history is crucial for unravelling the intricate
landscape of intratumour heterogeneity underlying cancer progression.
Several bioinformatics tools have been designed for deciphering the subclonal
population of the heterogeneous tumour mass. However, most of them rely on
single-omics analysis and methods for integrating multi-omics data in the
context of tumour evolutionary trees are still lacking.
In this thesis, the development of MAPping SubClonal Events (MAPSCE), a
new tool for mapping of subclonal events on tumour evolutionary trees, is
described. This method allows for integration of multi-omics data in multisample
cancer evolutionary studies. In essence, MAPSCE implements a
branch test where quadratic programming is applied to every branch of a
patient tumour tree to find the best mapping branch (including the root). Each
solution translates into a Bayesian Information Criterion value, and Bayes
factors for model selection. MAPSCE has been released as an R package.
Multiple datasets with different types of copy number events and varying
degrees of noise up to ±30% were simulated to assess the reliability of the
tool. For losses of haploid genes, MAPSCE was benchmarked against a tool
of similar functionality, LOHHLA, showing both an increase in specificity and
sensitivity. This comparison was not possible for other types of copy number
events as MAPSCE is the only tool to date with the ability to map these.
Lastly, MAPSCE’s potential applications were demonstrated in several
analyses of multi-region, multi-omics datasets. Subclonal biallelic inactivation
of tumour suppressor genes on subclonal level was identified in lung cancer
patients. Subclonal changes of gene expression were further compared
against subclonal copy number events to infer cases of copy number
dependent or independent allele specific expression.
This work provides an innovative way to integrate multi-omics data in multisample
cancer studies, refining the study of evolutionary processes underlying
intratumour heterogeneity.