Phylodynamics of Cell Populations - Human diseases are profoundly shaped by cell proliferation, differentiation, and movement among tissues. Proliferation of self-reactive immune cells cause autoimmune diseases, improper differentiation leads to developmental diseases, and spread of pathogenic bacteria among tissues are a part of infection and sepsis. Many drugs also work by perturbing cell populations: vaccines drive proliferation and differentiation of pathogen- targeting immune cells, and antibiotics cause bacterial cell death but also selection for antibiotic resistance. The fundamental mechanisms of these diseases and treatments cannot be understood without deciphering these dynamic cellular processes. While model organisms like mice can be studied using invasive experimental methods, humans can typically only be studied using blood and tissue samples. Even with the latest single cell sequencing technologies, these samples are small, static glimpses of cells in a particular time and tissue. New computational methods are needed that use single cell sequence data to decode the pathogenic mechanisms of diseases by reconstructing the history of cell proliferation, differentiation, and migration events. The goal of this proposal is to decode the mechanisms of human diseases by developing new computational models, inspired by evolutionary biology, that detect cellular proliferation, differentiation, and migration from single cell sequence data. Evolutionary biologists face similar problems as researchers studying humans – inferring population dynamics from small samples of sequences. For example, evolutionary models were used to trace the origin and spread of SARS-CoV-2 in 2020. However, these methods for viruses are inappropriate for cell populations due to their biased mutations and heterogenous evolutionary rates. We will address these challenges and develop phylodynamic models that use single cell sequencing data to infer recent changes in cell population size, patterns of differentiation over time, and patterns of cellular migration among tissues. The computational methods developed in this proposal will be broadly applicable across cellular systems and data types, and released as open source software. As input data, we will use B cell receptor sequences, bacterial genomes, and mitochondrial DNA mutations. As example applications, we will use these methods across cell types and species to i) determine the timing of B cell expansion and differentiation following influenza vaccination, ii) infer patterns of bacterial migration in response to antibiotic treatment in cystic fibrosis lung infections, and iii) identify changes in hematopoietic stem cell differentiation patterns with age. The ultimate vision of this proposal is to build advanced computational methods that use single cell sequence data to infer cellular dynamics underlying the pathology and treatment of human diseases.
