UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Copy-number aware methylation deconvolution analysis of cancers

Larose Cadieux, Elizabeth; (2021) Copy-number aware methylation deconvolution analysis of cancers. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of PhD_Thesis_ELC.pdf]
Preview
Text
PhD_Thesis_ELC.pdf

Download (33MB) | Preview

Abstract

DNA methylation has long been known to play a role in tumourigenesis. To this day, interpretation of bulk tumour bisulphite sequencing data has been hampered by normal contamination levels and tumour copy number. To address this issue, we develop two computational tools: (1) ASCAT.m, which allows Allele-Specific Copy number Analysis of Tumour methylation data directly from bulk tumour reduced representation bisulphite sequencing (RRBS) data and (2) CAMDAC, a method for Copy Number-Aware Methylation Deconvolution Analysis of Cancer, from bulk tumour and adjacent normal RRBS data. We describe a set of rules to compute allelic imbalance independently of bisulphite conversion and correct normalised read coverage estimates for sequencing biases. We apply ASCAT.m to non-small cell lung cancers from the epiTRACERx study with multi-region bulk tumour RRBS and adjacent normal. ASCAT.m genotypes, allele-specific copy numbers and tumour purity and ploidy estimates are in excellent agreement with those obtained from matched whole-exome and a subset of whole-genome sequencing of the same samples. We observe a correlation between whole-genome doubling and relapse-free survival in lung squamous cell carcinoma but not in adenocarcinoma. We see widespread genomic instability across both histological subtypes. We model bulk tumour methylation rates as a mixture of tumour and normal signals weighed for tumour purity and copy number and formalise this relationship into CAMDAC equations. The errors between predicted and observed methylation rates were low. Normal infiltrates Fluorescence-activated cell sorting (FACS)-purified from the bulk tumour were similar in composition to the adjacent matched normal lung, suggesting the latter is a suitable proxy for deconvolution. Single nucleotide variant (SNV)- and FACS-purified tumour methylation rates are in good agreement with CAMDAC deconvoluted estimates. Purification successfully removes shared normal signal, decreasing correlations between patients and to normal after purification. Samples with shared ancestry remain highly correlated. Purified methylation rates yield accurate tumour-normal and tumour-tumour differential methylation calls independent of tumour purity and copy number. We find hundreds of ubiquitously early clonal gene promoter epimutations across the epiTRACERx cohort, showcasing the potential of DNA methylation markers for early cancer detection. CAMDAC purified profiles reveal both phylogenetic and inter-tumour relationships as well as provide insight in tumour evolutionary history. Quantifying allele-specific methylation on chromosome X in females, we uncover extraction biases against the Barr body. X inactivation is random at the scale of our normal lung cancer samples. Phasing of methylation rates with polymorphisms confirms the presence of allele-specific methylation at the H19/IGF2 locus. Loss of imprinting is observed in 5 tumours, all involving demethylation of the maternal allele. We attempt to quantify the ratio of clonal allele-specific to bi-allelic epimutations in tumours in regions of 1+1, which we define as regulatory and stochastic methylation changes, respectively. Utilising this ratio, we try to extract the number of stochastic epimutations in regions of 2+0 with copy numbers 1 and 2 and time those copy number gains. We find that SNVs at gene promoters often lead to hypermethylation of neighbouring CpGs on the same copy or allele, suggesting the ablation of a transcription factor binding site. Non-expressed neo-antigen are enriched for promoter hypermethylation, indicating methylation plays a role in immune escape. To conclude, CAMDAC purified methylation rates are key to unlock insights into comparative cancer epigenomics and intra-tumour epigenetic heterogeneity.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Copy-number aware methylation deconvolution analysis of cancers
Event: UCL (University College London)
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Copyright © The Author. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request.
Keywords: Cancer epigenetics, Non-small cell lung cancer, DNA methylation, Methylation deconvolution
UCL classification: UCL
UCL > Provost and Vice Provost Offices
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences > Div of Biosciences
URI: https://discovery.ucl.ac.uk/id/eprint/10125248
Downloads since deposit
179Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item