Lineage reconstruction from clonal correlations

Abstract

<jats:p>A central task in developmental biology is to learn the sequence of fate decisions that leads to each mature cell type in a tissue or organism. Recently, clonal labeling of cells using DNA barcodes has emerged as a powerful approach for identifying cells that share a common ancestry of fate decisions. Here we explore the idea that stochasticity of cell fate choice during tissue development could be harnessed to read out lineage relationships after a single step of clonal barcoding. By considering a generalized multitype branching process, we determine the conditions under which the final distribution of barcodes over observed cell types encodes their bona fide lineage relationships. We then propose a method for inferring the order of fate decisions. Our theory predicts a set of symmetries of barcode covariance that serves as a consistency check for the validity of the method. We show that broken symmetries may be used to detect multiple paths of differentiation to the same cell types. We provide computational tools for general use. When applied to barcoding data in hematopoiesis, these tools reconstruct the classical hematopoietic hierarchy and detect couplings between monocytes and dendritic cells and between erythrocytes and basophils that suggest multiple pathways of differentiation for these lineages.</jats:p>

How Our Cells Become Our Selves: The Cellular Phylodynamic Biology of Growth and Development

“…Following the thinking of Stadler, Pybus, and Stumpf, 22 we call these "Cellular Population Tree Visualization Simulations", which should bring the powerful mathematics of lineage formation, coalescent theory, [22][23][24][25][26][27] to developmental biology. For example, there has been much interest in the past few years in extracting information on gene expression, differentiation, and lineage formation, from single cell RNA expression, 50,53,54 cellular barcoding, 21,51 and somatic mutation. 52 However, as Telford 21 and Klein 53,54 have pointed out, without knowledge of the mitotic pattern of an animal's cell lineage, extraction of useful information from such single cell data is problematic.…”

“…For example, there has been much interest in the past few years in extracting information on gene expression, differentiation, and lineage formation, from single cell RNA expression, 50,53,54 cellular barcoding, 21,51 and somatic mutation. 52 However, as Telford 21 and Klein 53,54 have pointed out, without knowledge of the mitotic pattern of an animal's cell lineage, extraction of useful information from such single cell data is problematic. One might hope, and indeed expect, that the potential of Cellular Population Tree Visualization Simulation, driven by the parameters measured by Cellular Phylodynamic Analysis, could unburden single cell analysis from the requirement of deciphering ancestry, and unlock the much hoped-for treasure of how gene expression is linked to cell lineage formation.…”

“…15,16 Cell lineages grow from individual Founder Cells, 2,17 although how embryos use mitosis to create these parts, 18,19 and how these Founder Cells undergo Cell-Heritable change that characterized each clone, 16,35 remains unknown. The analysis of cellular diversification, by single cell mRNA expression, 50,53,54 cell marking, 21,20,51 and high resolution 4D microscopy, 18,19,47 has also wrestled with the problem of cell lineage formation. 21 Here, we report the examination of growth and development in discrete units of numbers of cells, N. We call this approach "Cellular Phylodynamics".…”

Learning dynamics by computational integration of single cell genomic and lineage information

“… c , The normalized covariance of clonal barcode abundances between different cell types, calculated using all data on day 4 of differentiation 1 .…”

A Unified Framework for Lineage Tracing and Trajectory Inference

“…High-resolution lineage tracing, however, does not obviate the need for trajectory inference because lineage tracing alone does not reveal the state of the ancestral cells. While the problems of reconstructing lineage trees and inferring trajectories have attracted substantial attention individually [20, 33], there is much to be gained from combining these two complementary perspectives [6].…”