Inferring transcriptional bursting kinetics from single-cell snapshot data using a generalized telegraph model

Abstract

Gene expression has inherent stochasticity resulting from transcription's burst manners. Single-cell snapshot data can be exploited to rigorously infer transcriptional burst kinetics, using mathematical models as blueprints. The classical telegraph model (CTM) has been widely used to explain transcriptional bursting with Markovian assumptions. However, growing evidence suggests that the gene-state dwell times are generally non-exponential, as gene-state switching is a multi-step process in organisms. Therefore, interpretable non-Markovian mathematical models and efficient statistical inference methods are urgently required in investigating transcriptional burst kinetics. We develop an interpretable and tractable model, the generalized telegraph model (GTM), to characterize transcriptional bursting that allows arbitrary dwell-time distributions, rather than exponential distributions, to be incorporated into the ON and OFF switching process. Based on the GTM, we propose an inference method for transcriptional bursting kinetics using an approximate Bayesian computation framework. This method demonstrates an efficient and scalable estimation of burst frequency and burst size on synthetic data. Further, the application of inference to genome-wide data from mouse embryonic fibroblasts reveals that GTM would estimate lower burst frequency and higher burst size than those estimated by CTM. In conclusion, the GTM and the corresponding inference method are effective tools to infer dynamic transcriptional bursting from static single-cell snapshot data.

Global transcription regulation revealed from dynamical correlations in time-resolved single-cell RNA-sequencing

“…The method allows quantifying distributions of transcript levels across cells that can be harnessed to learn about gene regulation. A few studies have started to investigate how transcriptional processes are regulated through synergising stochastic modelling and inference [31][32][33][34][35]. To this end, stochastic models of reaction kinetics inside living cells are simulated using the Gillespie algorithm or solved analytically and matched with experimental mRNA counts using statistical inference [24,[36][37][38][39][40].…”

“…In the literature, these problems are typically dealt with separately. The first issue is addressed by replacing likelihoods with momentor simulation-based methods [33, 35, 37, 43]. Yet, these methods remain computationally challenging as they need to be performed for several competing models and many genes.…”

“…In the literature, these problems are typically dealt with separately. The first issue is addressed by replacing likelihoods with moment-or simulation-based methods [33][34][35][36]44]. As an example, the authors of [34] present an Approximate Bayesian Computation framework to infer transcriptional burst kinetics from scRNA-seq data.…”

“…The first issue is addressed by replacing likelihoods with moment-or simulation-based methods [33][34][35][36]44]. As an example, the authors of [34] present an Approximate Bayesian Computation framework to infer transcriptional burst kinetics from scRNA-seq data. Yet, the above methods either neglect technical noise of sequencing protocols or remain computationally challenging as they need to be performed for several competing models and many genes.…”

Nascent RNA kinetics with complex promoter architecture: Analytic results and parameter inference

“…An intriguing question is whether the memory of promoter switching (called promoter memory for convenience) can be inferred from the stationary nascent RNA distribution. The question of how model parameters are inferred from a given distribution has been explored in previous studies [47-50]; and the two-state non-Markov model has commonly been employed to fit the nascent RNA distribution and to elucidate the underlying mechanism of gene expression [14,15]. In this section, we propose a method based on the above analytical results to infer the promoter memory from the stationary nascent RNA distribution.…”