Figure 1. Overview of statistical computing framework for in silico phenotype projection. (a) experimental perturbation of ERK1/2 in human umbilical vein endothelial cells (HUVEC), data processing, feature learning, deep learning, and statistical assessment of model classification performance of single cell RNA-seq data. (b) Bayesian Belief Network analysis (BBN), experimental validation of putative causal gene dependency structure in HUVEC, experimental evaluation of predicted phenotypes in mice, and natural language processing of results. Modified from Figure S5, Ricard et al., JEM 2019.