RNA translation to protein is paramount to creating life, yet RNA and protein correlations vary widely across tissues, cells, and species. To investigate these perplexing results, we utilize a time-series fixation method that combines static stimulation and a programmable formaldehyde perfusion to map pseudo-Signaling with Omics signatures (pSigOmics) of single-cell data from hundreds of thousands of cells. Using the widely studied nuclear factor kappa B (NFκB) mammalian signaling pathway in mouse fibroblasts, we discovered a novel asynchronous pseudotime regulation (APR) between RNA and protein levels in the quintessential NFκB p65 protein using single molecule spatial imaging. Prototypical NFκB dynamics are successfully confirmed by the rise and fall of NFκB response as well as A20 negative inhibitor activity by 90 min. The observed p65 translational APR is evident in both statically sampled timepoints and dynamic response gradients from programmable formaldehyde fixation, which successfully creates continuous response measurements. Finally, we implement a graph neural network model capable of predicting APR cell subpopulations from GAPDH RNA spatial expression, which is strongly correlated with p65 RNA signatures. Successful decision tree classifiers on Potential of Heat-diffusion for Affinity-based Trajectory Embedding embeddings of our data, which illustrate partitions of APR cell subpopulations in latent space, further confirm the APR patterns. Together, our data suggest an RNA-protein regulatory framework in which translation adapts to signaling events and illuminates how immune signaling is timed across various cell subpopulations.
Read full abstract