Stochastic mRNA synthesis in mammalian cells was investigated by counting individual mRNA molecules from a reporter gene stably integrated into the genome. The study revealed massive variations in mRNA numbers per cell, attributed to bursts of transcription during gene activation and inactivation. These transitions are intrinsic and not due to extrinsic factors like transcriptional activators. Burst-like expression affects entire genomic loci and is observed in natural genes. mRNA bursts can be buffered by slow protein degradation. A stochastic model of gene activation and inactivation explained the statistical properties of bursts, showing that increasing transcription factors increases burst size, not frequency. The results indicate that gene expression in mammalian cells is subject to large, intrinsic fluctuations, raising questions about how cells function in the face of such noise. The study used fluorescence in situ hybridization (FISH) to detect single mRNA molecules, revealing that gene activation and inactivation events are infrequent and random. The findings suggest that variations in mRNA levels are intrinsic and not due to global factors. The study also showed that the mRNA encoding the large subunit of RNA polymerase II is burst-like and not correlated with reporter gene mRNA. Protein levels were found to be less affected by mRNA variability when degradation rates are slow. The results highlight the importance of intrinsic randomness in gene expression and its implications for cellular function. The study used mathematical models to analyze the distribution of mRNA and protein levels, showing that burst-like transcription can lead to skewed distributions. The findings have implications for understanding gene regulation in both unicellular and multicellular organisms.Stochastic mRNA synthesis in mammalian cells was investigated by counting individual mRNA molecules from a reporter gene stably integrated into the genome. The study revealed massive variations in mRNA numbers per cell, attributed to bursts of transcription during gene activation and inactivation. These transitions are intrinsic and not due to extrinsic factors like transcriptional activators. Burst-like expression affects entire genomic loci and is observed in natural genes. mRNA bursts can be buffered by slow protein degradation. A stochastic model of gene activation and inactivation explained the statistical properties of bursts, showing that increasing transcription factors increases burst size, not frequency. The results indicate that gene expression in mammalian cells is subject to large, intrinsic fluctuations, raising questions about how cells function in the face of such noise. The study used fluorescence in situ hybridization (FISH) to detect single mRNA molecules, revealing that gene activation and inactivation events are infrequent and random. The findings suggest that variations in mRNA levels are intrinsic and not due to global factors. The study also showed that the mRNA encoding the large subunit of RNA polymerase II is burst-like and not correlated with reporter gene mRNA. Protein levels were found to be less affected by mRNA variability when degradation rates are slow. The results highlight the importance of intrinsic randomness in gene expression and its implications for cellular function. The study used mathematical models to analyze the distribution of mRNA and protein levels, showing that burst-like transcription can lead to skewed distributions. The findings have implications for understanding gene regulation in both unicellular and multicellular organisms.