Whole-organisms lineage tracing inferred by genotyping of induced somatic mutations: hopes and hypes

Complex animals consist of millions of cells that derive from a single cell, the zygote. A long-standing aim in developmental biology has been to unravel an organism’s full cell lineage, i.e., how cells relate to each other. Several research groups have recently developed CRISPR/Cas9-based lineage recorders, which accumulate mutations as cells divide during development. These mutations are then used retrospectively to infer the lineage relationships between cells. CRISPR-based lineage recording methods are potentially scalable to whole-organism, with a single-cell resolution, and have produced high expectations in the developmental biology field. Nevertheless, while the basic principle of recorder-based lineage tree reconstruction is clear and intuitive, the accuracy of the lineages inferred by these methods has not been systematically evaluated. It is evident that the conclusions we draw will only be useful if the tree is accurate. The ability to estimate the relative accuracy of different methods would also be vital when trying to optimize alternative approaches. Our purpose is to quantify how accurate existing methods can be when reconstructing a cell lineage of tens of thousands of cells, trying to highlight some of the potential pitfalls when facing experimental constraints. To this purpose, we use extensive simulations (informed by real experimental data) to quantify how different factors (recorder design, mutation rates, cell lineage depth, etc.) can undermine the accuracy of lineage reconstruction. Finally, we propose a set of criteria for the optimal design of CRISPR-based lineage recorders, as well as some guidelines to understand their potential limitations.