For differential gene expression analyses, gene expression estimates (TPM) were log-transformed after the addition of a constant (1) to each expression estimate. The top 58% (27,446 genes) most highly expressed genes across cell types (mean expression in iPSCs, HLCs, primary hepatocytes, and whole livers) were used for genome-wide PCA using the pcaMethods R package (Stacklies et al., 2007).