Using the simple telegraph model to decipher transcriptional burst regulation across genome-wide data

Gene transcription is a stochastic bursting process, with burst frequency and burst size as core parameters. Deciphering its genome-wide regulation is challenging, as complex models (capturing detailed biology) are impractical for large-scale use due to unstable inference and high computation. We proposed a simple telegraph model-based framework to analyze genome-wide scRNA-seq and smFISH datasets, addressing these limitations. When sample size ≥ 500 and burst parameter change ≥ 3-fold, its inferred burst frequency- and size-dominated variations reliably proxy true regulation. Analyses across mouse cells (CAST/C57 alleles, fibroblasts vs. embryonic stem cells) and healthy/hypertrophic cardiomyopathy (HCM) human heart tissues revealed key rules: 1) Burst frequency- or size-dominated regulation was primary (over 70% of variable genes), with more bursty-regulated genes in HCM; 2) TATA-initiator synergy (enhancing burst size-dominated regulation) was lost in HCM; 3) Burst frequency-dominated genes enriched in genome stability/cell cycle/apoptosis (via TFs like Foxo4 / Mcm2 ), while burst size-dominated ones in signaling/metabolism/proteostasis (via TFs like Zfp322a / Smchd1 / Ppargc1a ), and their interdependent dysregulation accelerated HCM. This study establishes the simple telegraph model as a scalable framework linking transcriptional burst dynamics to cell fate and pathology.