2Gene expression programs are dynamic, e.g. the cell cycle, response to stimuli, normal differentiation and 3 development, etc. However, nearly all techniques for profiling gene expression in single cells fail to 4 directly capture the dynamics of transcriptional programs, which limits the scope of biology that can be 5 effectively investigated. Towards addressing this, we developed sci-fate, a new technique that combines 6 S4U labeling of newly synthesized mRNA with single cell combinatorial indexing (sci-), in order to 7 concurrently profile the whole and newly synthesized transcriptome in each of many single cells. As a 8 proof-of-concept, we applied sci-fate to a model system of cortisol response and characterized expression 9 dynamics in over 6,000 single cells. From these data, we quantify the dynamics of the cell cycle and 10 glucocorticoid receptor activation, while also exploring their intersection. We furthermore use these data 11 to develop a framework for inferring the distribution of cell state transitions. We anticipate sci-fate will 12 be broadly applicable to quantitatively characterize transcriptional dynamics in diverse systems. 13 Main 1 2During organismal development, as well as during myriad physiological and pathophysiological 3 processes, individual cells traverse a manifold of molecularly and functionally distinct states. The accurate 4 characterization of these trajectories is key to advancing our understanding of each such process, for 5 identifying the causal factors that drive them, and for rationally designing effective perturbations of them. 6 However, although experimental methods for profiling various aspects of single cell biology have recently 7 proliferated, nearly all such methods deliver only a static snapshot of each cell, e.g. of gene expression at 8 the moment of fixation. 9 10To recover temporal dynamics, several groups have developed computational methods that place 11 individual cells along a continuous trajectory based on single cell RNA-seq data, i.e. the concept of 12 pseudotime 1-6 . However, such methods are inherently limited in several important ways, including that 13 they are inferring rather than directly measuring dynamics, that they are dependent on sufficient 14 representation across the trajectory, and that they may fail to capture the detailed dynamics of individual 15 cells (e.g. directionality, multiple superimposed potentials, etc.) 7 . Although time-lapse microscopy is a 16 distinct technology that overcomes some of these limitations, it is limited in throughput and scope (e.g.
17enabling visualization of a few marker genes in a few cells), and as such may be insufficient to decipher 18 the complexity of many biological systems.
20Here we describe a novel technique, sci-fate, to measure the dynamics of gene expression in single cells 21 at the level of the whole transcriptome. In brief, we integrated protocols for labeling newly synthesized 22 mRNA with 4-thiouridine (S4U) 8,9 with single cell combinatorial indexing RNA-seq (sci-RNA-seq 10 ). As 23 a p...