BoolODE Simulated Time Series within Single Cells

BoolODE was designed to simulate single-cell expression data sets on the basis of a network of gene-gene interactions (Pratapa et al., 2020). We applied it here to generate expression simulated expression levels at 800 time points within the same cell in order to benchmark predictions of future expression levels. To generate the different simulations (117), we first needed different gene-gene networks. These were created as random ten gene networks, where genes could have positive, negative, or no direct relationship with other genes. Each network was input to BoolODE, which simulates hundreds of time points of expression for 2000 cells in each simulation. Minor changes were made to BoolODE code, to generate output for the task of predicting future expression states.