Abstract 239: Transcriptomic Profiling Maps Anatomically Patterned Subpopulations Among Single Embryonic Cardiac Cells

Abstract

Embryonic gene expression intricately reflects anatomical context, developmental stage, and cell type. To address whether the precise spatial origins of cardiac cells can be deduced solely from their transcriptional profiles, we established a genomewide expression database from 118, 949, and 1,166 single murine heart cells at embryonic day 8.5 (e8.5), e9.5, and e10.5, respectively. We segregated these cells by type using unsupervised bioinformatics analysis and identified chamber-specific genes. Using a random forest algorithm, we reconstructed the spatial origin of single e9.5 and e10.5 cardiomyocytes with 92.0% ± 3.2% and 91.2% ± 2.8% accuracy, respectively (99.4% ± 1.0% and 99.1% ± 1.1% if a ±1 zone margin is permitted) and predicted the second heart field distribution of Isl-1-lineage descendants. When applied to Nkx2-5/ cardiomyocytes from murine e9.5 hearts, we showed their transcriptional alteration and lack of ventricular phenotype. Our database and zone classification algorithm will enable the discovery of novel mechanisms in early cardiac development and disease.