Characterization of human pluripotent stem cell differentiation by single-cell dual-omics analyses (scATAC-seq)

In vivo differentiation of human pluripotent stem cells (hPSCs) has unique advantages such as multilineage differentiation, angiogenesis, and close cell?cell interactions. To systematically investigate the multilineage differentiation mechanisms of hPSCs, we constructed the in vivo hPSC differentiation landscape containing 239,670 cells using teratoma models. We inferred 18 cell differentiation trajectories and characterized common and specific gene regulation patterns during hPSC differentiation at both transcriptional and epigenetic levels. The results of developing single cell Basic Local Alignment Search Tool (dscBLAST), our self-developed tool for cell identification and developmental stage speculation, further confirmed the differentiated cell identity and suggested that cells in the teratoma are largely at immature developmental stages. Overall, our study offers new insights into stem cell fate decisions and embryonic development; dscBLAST shows favorable cell identification performance, providing both developmental and adult cells with a powerful tool for cell annotation and stage speculation. Overall design: To construct the landscape of teratoma,18 teratomas were used for single-cell RNA sequencing (scRNA-seq) based on Microwell-seq. Among them, six were generated from H9 embryonic stem cells (ESCs), and 12 were from H1 ESCs. To conduct joint analysis of dual-omics data, additional three teratomas from H9 ESCs were utilized for scRNA-seq and scATAC-seq simultaneously. For each teratoma sample, there are human-derived cells differentiated from ESCs and mouse-derived cells infiltrating from the thigh muscles of NOD/SCID mice, where the ESCs were injected.