Speaker
Description
Live-cell imaging experiments often track mRNA production using fluorescently tagged transcripts. To infer the statistics of initiation events, standard methods assume that transcription has a fixed duration. However, studies of intron splicing and transcriptional pausing show that stochasticity in elongation and termination are often significant and cannot be ignored. In this talk, we advocate an alternative approach: fitting directly to the autocorrelation function of the fluorescence intensity signal. Building on recent links between transcriptional models and infinite‑server queueing systems, we use theoretical autocorrelation expressions to infer parameters in models that incorporate stochasticity in both gene switching and transcription time. This framework provides richer mechanistic insight than approaches requiring fixed transcription durations. We illustrate its advantages using recent live‑cell imaging data, fitting autocorrelations under a delay Telegraph model that includes an additional exponential waiting time in the transcription process.
