Cell lineages provide the framework for understanding how multicellular organisms are built and how cell fates are decided during development. Describing cell lineages in most organisms is challenging, given the number of cells involved; even a fruit fly larva has ∼50,000 cells and a small mammal has more than 1 billion cells. Recently, the idea of using CRISPR to induce mutations during development as heritable markers for lineage reconstruction has been proposed and trialled by several groups. While an attractive idea, its practical value depends on the accuracy of the cell lineages that can be generated by this method. Here, we use computer simulations to estimate the performance of this approach under different conditions. Our simulations incorporate empirical data on CRISPR-induced mutation frequencies in Drosophila. We show significant impacts from multiple biological and technical parameters -variable cell division rates, skewed mutational outcomes, target dropouts and different mutation sequencing strategies. Our approach reveals the limitations of recently published CRISPR recorders, and indicates how future implementations can be optimised to produce accurate cell lineages.
Correspondence:michalis.averof@ens-lyon.fr m.telford@ucl.ac.uk
IntroductionStarting from a single cell -the fertilised egg -multicellular organisms undergo repeated rounds of cell division to produce the adult form. The divisions that generate these adult cells constitute a genealogical tree with the fertilised egg at its root and each differentiated cell as a terminal branch. Knowing the cell lineage that produces a fully developed organism from a single cell provides the framework for understanding when, where and how cell fate decisions are made. Obtaining high resolution (single-cell level) lineages is a challenging task that has been solved only in simple cases, such as the nematode Caenorhabditis elegans: its complete lineage (∼1000 cells) was deduced by painstaking observation of each cell division under the microscope. This approach is impossible in larger animals, in which most cells are inaccessible to microscopy and their number becomes quickly unmanageable. The 16 rounds of cell division required to produce a hatched Drosophila larva, for example, result in about 50,000 cells (1) and further rounds of division produce an adult with approximately 10 6 cells. The bodies of mice and humans consist of 10 10 to 10 14 cells respectively Left: Development begins with a zygote carrying in its genome a lineage recorder composed of a series of CRISPR targets (blue boxes). During subsequent cell divisions, any target of the recorder can be cleaved by Cas9 in any cell, leaving a specific mutational signature on the target which will be inherited by all the descendants of the cell. Numbers represent the the cleaved target in the recorder and its mutational signature is represented with a colour. Middle: At the end of development, the recorder of every cell is sequenced, recovering the pattern of accumulated mutations in each of the t...