next on phyloseminar.org
To attend a seminar, please visit the livestream portion of our YouTube channel.
Phylogenetics of Cancer 2
Modeling tumor progression from single-cell sequencing data
Cancer progression is a somatic evolutionary process characterized by the accumulation of genetic alterations and responsible for tumor growth, clinical progression, and drug resistance development. I will present probabilistic models for reconstructing the evolutionary history of a tumor from single-cell sequencing data, including efficient inference algorithms for mutation calling and for learning tumor phylogenies from mutation and copy number data. I will also discuss computational methods for finding common patterns of tumor evolution among patients and for inferring the underlying cellular fitness landscape of cancer evolution.
Tumor Evolution over Space and Time
Cancer is the result of an evolutionary process where genomic mutations and epigenetic alterations accumulate in cells that form a tumor. Modern single-cell and spatial sequencing technologies enable the measurement of mutations at thousands of cells or spatial locations within a tumor. However, due to technical limitations these measurements are typically sparse with high rates of missing data. In this talk, I will describe algorithms that overcome these limitations and: (1) Reconstruct tumor evolution in space from large copy number aberrations, revealing directions of tumor growth in 2D and 3D; (2) Analyze tumor evolution over time and in response to treatment using longitudinal single-cell sequencing.
Quantitative measurement of cancer evolutionary history from single-sample bulk DNA methylation data
Cancers evolve. Current methods to measure evolutionary dynamics rely on genome sequencing of multiple samples from each tumour to detect subclonal mutations and using them to perform phylogenetic tree inference. In this talk I will show a very different approach that uses only single-sample bulk DNA methylation data to provide an unexpectedly rich read out of the clonal history of tumour evolution. Our approach relies on “fluctuating DNA methylation” - the random loss and gain of methylation at particular CpG sites which are evolutionarily neutral. We infer evolutionary history by modelling the site frequency spectrum of fluctuating methylation within single samples. I will show applications in haematological cancers where we use the approach to measure, in 1000s of cancers, tumour age, growth rate and various features of subclonal dynamics.