University of British Columbia
LiquidBayes : a bayesian network for monitoring cancer progression using liquid biopsies
Abstract
dc:descriptionCancer exhibits temporal heterogeneity, described by the existence and continual evolution of multiple cell subpopulations in a tumour. Temporal heterogeneity triggers resistance, disrupting targeted therapies and worsening patient prognosis. The continual monitoring of cancer patients can aid in identifying resistance and lead to informed treatment decisions. Liquid biopsies are non-invasive blood samples containing double stranded, free-floating DNA called cell-free DNA (cfDNA). A subset of cfDNA, known as circulating tumour DNA (ctDNA), originates from the tumour itself and can uncover tumour-specific mutations to characterize cancer. Although liquid biopsies provide a means to evaluate response over time, ctDNA abundance can be extremely low in post-treatment patients, inhibiting its utility for monitoring therapy efficacy in these contexts. Statistical methods have been developed for estimating tumour fraction using ctDNA samples. Moreover, some groups have integrated somatic mutations derived from bulk sequencing of a tissue biopsy to address low ctDNA abundance. However, no method we know of incorporates single-cell sequencing of a tissue biopsy in ctDNA analysis. In this thesis, we present LiquidBayes, a Bayesian Network that integrates clone-level copy number profiles from single-cell Whole Genome Sequencing (WGS) of the primary tissue with WGS of ctDNA samples. LiquidBayes leverages Markov Chain Monte Carlo (MCMC) techniques and is implemented using a Probabilistic Programming Language (PPL). LiquidBayes significantly outperforms state-of-the-art methods in tumour fraction estimation and allows for inference of clonal prevalences. LiquidBayes can analyze serial ctDNA samples to dissect temporal heterogeneity, intercept resistance and ultimately improve patient prognosis.
Degree
thesis:*- Name thesis:degree_name
- Master of Science - MSc
- Level thesis:degree_level
- master's
- Discipline thesis:degree_discipline
- Bioinformatics
- Grantor dc:publisher
- University of British Columbia
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Yang, Kevin
- Contributors dc:contributor
-
- Roth, Andrew J. L.
- Park, Yongjin P.
Rights
dc:rights- Statement dc:rights
-
- Attribution-NonCommercial-NoDerivatives 4.0 International
- Language dc:language
- eng
Identifiers
dc:identifier.*- Handle dc:identifier
- http://hdl.handle.net/2429/83469