Biddy, B. A. et al. Nature 564, 219–224 (2018).

Direct reprogramming converts one somatic cell type to another, but the process is inefficient, which makes it difficult to study the molecular mechanism of lineage conversion. Biddy et al. use a lentiviral-based approach to label cells with barcode combinations that record both cell identity and clonal history. The researchers applied their ‘CellTagging’ to mouse embryonic fibroblasts, which they reprogrammed to endoderm progenitors (iEPs). After sequencing the tags and transcriptomes of over 80,000 cells, clustering them, and tracking iEP markers, they saw that many cells had initiated reprogramming but few completed the transition. They found that iEPs were derived from only a few clones, and lineage reconstruction showed a split in the trajectories, with one set of clones resulting in iEPs and the others resulting in a ‘dead end’. This split occurred on day 6 of reprogramming, and expression of the methyltransferase Mettl17a1 was upregulated in the successfully reprogrammed clones. Knowing the contributors to direct reprogramming will help make the process more efficient.