Simulations of gene regulatory networks with transcriptional adaptation

Background

Cells and tissues have a remarkable ability to adapt to genetic perturbations via a variety of molecular mechanisms. Transcriptional adaptation has recently emerged as one such mechanism, in which nonsense mutations in a gene trigger upregulation of related genes, possibly conferring robustness at cellular and organismal levels. However, beyond a handful of developmental contexts and curated sets of genes, no comprehensive genome-wide investigation of this behavior has been undertaken for mammalian cell types and contexts. Further, how the regulatory-level effects of inherently stochastic compensatory gene networks contribute to phenotypic penetrance in single cells remains unclear.

Results

In the corresponding manuscript, we analyze existing bulk and single-cell transcriptomic datasets to infer the prevalence of transcriptional adaptation in mammalian systems across diverse contexts and cell types. In the data presented here, stochastic mathematical modeling of minimal compensatory gene networks qualitatively recapitulates several aspects of transcriptional adaptation. Combined with machine learning analysis of network features of interest, our framework offers potential explanations for which regulatory steps are most important for transcriptional adaptation.

Conclusions

We provide a formal quantitative framework to test and refine models of transcriptional adaptation.

Methods
The dataset was generated through simulations of gene regulatory networks with transcriptional adaptation, where genes are represented as nodes and regulatory relationships as edges. Please see the corresponding manuscript for a complete description of the model, simulation conditions, and analyses.

The regulatory networks were constructed using a modified telegraph model for transcriptional bursting, where each gene's alleles switch between active (transcribing) and inactive (quiescent) states, for an ancestral regulator gene (A), its paralogs (A'1 and A'2), and a downstream target gene (B). To capture the dynamics of gene regulation, we implemented stochastic simulations using Gillespie’s next reaction method. We accounted for differential regulatory effects of different gene products. We included models for both activating and repressing interactions, as well as models with more than one paralog gene. 

Parameter ranges were defined based on previously reported literature on transcriptional regulation and other related phenomena, as described in the corresponding manuscript. Parameter sets for simulations were sampled with latin hypercube sampling from the defined parameter ranges. Each simulation ran for 300,000 timesteps, simulating three genotypes: wildtype, heterozygous, and homozygous-mutant for 100,000 timesteps each. A total of approximately 60,000 simulations were performed across different network models. 

The simulations were performed in MATLAB R2017a, R2021b, and R2024a. This extensive dataset and the corresponding analysis provides valuable insights into the stochastic behavior of gene regulatory networks with transcriptional adaptation.

For a full description of the models and simulations, please see the preprint and corresponding forthcoming publication.