Bioinformatics for single-cell genome sequence analyses to study genome instability and intra-tumour genetic heterogeneity at high resolution

Although all cells in a human body are descendant from a single cell –i.e. the zygote– the genetic content in the different cells is not necessarily identical due to the accumulation of mutations during development and aging, making every individual a genetic mosaic. Such mutations may lead to the development of disease, like cancer or developmental disorders. Hence, it is important to study the spectrum of mutations that accumulate in a human lifetime, their nature and impact in different cell types. The failure of conventional methods to detect the underlying heterogeneity to the biological unit leads to the requirement of a novel technology. Single-cell sequencing solves this problem enabling the detection of rare subpopulations and the mutations that co-occur in the same single cell. To enable single-cell sequencing, the genetic content present in the cell needs to be amplified using whole genome amplification (WGA) methods at the expense of artefacts introduced during WGA that may resemble real genetic variants. At the commencement of my PhD in 2011, single-cell sequencing was still in its infancy and the assessment of different whole genome amplification methods for single-cell analysis was a need of the hour. In chapter 3 of this thesis, we applied two different WGA methods based on PCR and multiple displacement amplification (MDA) to provide the first genome-wide paired-end sequencing maps of single cells and enabled the detection of genuine structural variants by filtering WGA errors in single cells. In follow-up experiments described in chapter 4, we tested eight different WGA methods that are based on either PCR, MDA or a combination of displacement pre-amplification and PCR (DA-PCR) in microlitre volumes and also evaluated MDA in reduced nanolitre volumes for copy number and single nucleotide variation (SNV) detection in single cells. We observed that PCR and DA-PCR methods are most suitable for DNA copy number profiling while MDA methods are preferred for SNV detection. However, MDA in nanolitre volumes could ameliorate DNA copy number profiling of single cells. To enable genotype-phenotype correlations at single-cell resolution, methods that could interrogate both the genome and the transcriptome of the same single cell were paramount. In chapter 5, we contributed to the first generation of technologies for multi-omics single-cell sequencing and developed G&T-seq, a method that allows sequencing the genome and transcriptome of the same single cell in parallel. This method enabled investigating how genomic variation across individual cells may be coupled with specific transcriptomic cell states. Copy number aberrations (CNAs) detected at the DNA level were correlated with the gene expression dosage at the RNA level in same single cell. Also, structural variants and single nucleotide variants identified in the genome sequences could be interrogated in the transcriptome data of the same single cell. Cancer is characterised by cellular heterogeneity and often comprises several populations of genetically aberrant cells. In chapter 6, we isolated single cells from 3 different sectors of a primary breast tumour to dissect the subclonal architecture and disclose mechanisms of tumour genome evolution in a spatial context. Understanding the biology of cellular heterogeneity in cancer is paramount to design novel cancer treatments. We not only disentangled the clonal architecture of the primary tumour, but considerably refined the phylogenetic tree built from bulk sequencing data disclosing the archaeological record of acquired DNA mutations in the cancer. Along with a gradual evolution of DNA imbalances and a presumably sudden chromothripsis event in the cancer cells, we also observed identical and similar aberrations occurring in spatially different subpopulations, suggesting independent but convergent tumour evolution within the primary tumour. In chapter 7 and 8, we isolated disseminated tumour cells (DTCs) from bone marrow aspirates of non-metastatic breast cancer patients on the basis of immunocytochemical and morphologic markers and studied those cells for the first time using single cell sequencing. Analysis of the cells’ CNA profiles revealed that only a quarter of the single cells are in fact genuine DTCs disseminating from the observed tumor. The remaining cells represented non-aberrant ‘normal’ cells and ‘aberrant cells of unknown origin’ that contained CNAs unrelated with the tumor. Additionally, the single-cell sequences of the DTCs helped us to further resolve the phylogenetic tree structure constructed with bulk DNA sequencing of the primary tumour. Lastly, our results showed that DTCs disseminate from the primary tumour in later stages of tumour evolution. In conclusion, single-cell sequencing is a powerful approach to study cellular heterogeneity in normal as well as diseased tissues and to understand their biology in normal development and disease progression.status: publishe