The LeDucq CHECKPOINT ATHERO aims to characterize the cell type specific immune checkpoint (IC) expression profiles during atherogenesis. Transcriptomics of atherosclerotic plaques helps detail mechanistic IC signaling pathways. Recently, single-cell RNA-sequencing (scRNA-seq) has become an integral part of cardiovascular research and superseded bulkRNA-seq to gain insight intoa single-cell resolution. These two techniques are closely related: scRNA-seq data retains cell-specific barcode information. Removing this barcode effectively converts it into bulkRNA-seq. The reserve conversion–from bulkRNA-seq back to scRNA-seq remains nearly impossible. In this project, they develop and train a machine-learning model using a Gaussian-mixture variational autoencoder for converting bulkRNA-seq data into synthetic scRNA-seq data. We will use bulk and scRNA-seq data from the aortas of a 30-week old ApoE-/- mice cohort to train and test the model to generate representative scRNA-seq data. This model will be applied to existing RNA-seq data from a time-course atherosclerosis study to gain single-cell level insights.