Alladi Ramakrishnan Hall
Model-based Bayesian inference of single-cell behaviors
Manikandan Narayanan
NIAID, National Institutes of Health, Bethesda, MD, USA
Different cells could make different amounts of biomolecules such as
gene mRNAs. New technologies are just emerging to measure the level of
many genes in single cells, but accurate quantification of biological
cell-to-cell differences is challenging due to the low level of
expression of many genes and substantial measurement noise. Here we
present a Bayesian approach to robustly quantify biological
cell-to-cell difference by integrating different types of data, viz.,
measurements directly obtained from single cells and those from random
pools of k-cells (e.g., k=10). Simulations and real data reveal
distinct scenarios where we may be able to reap benefits with such an
integrative approach, and application to human immune cell data
uncovers how immune activation can shape not only the average
expression level of a gene, but also its variability from cell to
cell.
This work is supported by the intramural program of NIH.
Done