Manu Setty: Computational modeling of lineage decision mechanisms using single-cell data

North West Seminar Series of Mathematical Biology and Data Science

Wednesday, 22nd March 2021 (hosted by Mudassar Iqbal)

Abstract:
How a cell decides its lineage is a long-standing and enigmatic question in molecular biology. Single-cell technologies offer unprecedented scale and resolution to study the cellular differentiation process and enable the alignment of cells into a unified trajectory representing their developmental progression. I will describe Palantir, the first single-cell trajectory detection algorithm to view the differentiation process as a probabilistic change in cell potential. Palantir models differentiation as a Markov process to compute the lineage biases and differentiation potential, a measure of plasticity, for all cells. Palantir modeling enables an accurate identification of lineage decision regions and associated gene expression and regulatory dynamics. I will then describe our study to characterize the spatial and temporal trajectories of mouse endoderm development. Combining Palantir modeling with a large-scale single-cell data, we uncovered an unexpected plasticity of embryonic epiblast cells to differentiate to the surrounding extra-embryonic endoderm at implantation and the concomitant emergence of the spatial signal in the endoderm. Our analysis further demonstrated the convergence of embryonic definitive and extra-embryonic visceral endoderm cells into spatially organized organ territories in the gut-tube at mid-gestation.