How many genes can CRISPR edit to engineer complex adaptations?

Polygenic traits require the coordinated effects of multiple genes. Such complex traits have been a long-term target of study for geneticists, but multiplex CRISPR--the editing of multiple loci in the genome via multiple guide RNAs--is in its infancy. Reviewing 106 plant studies using multiplex CRISPR, we find that the multiplexing capacity has doubled every 5.4 years. Furthermore, a systematic experiment with 8, 16, and 24 simultaneous targets in Arabidopsis thaliana reveals efficiency of up to 75% in 24-plex editing in transformed plants confirmed by sequencing. We surprisingly found that the level of multiplexing, or the number of the targets, causes lesser efficiency reductions than other uncontrolled factors such as gRNA design or natural variation across plants. In fact, mathematical modeling of the decay in editing efficiency as a function of the gRNA numbers, showed a logistic model with sustained efficiency fits the data better than decay due to Cas9 competition or random editing. We then project that editing close to 100 genes in a plant can be feasible with reasonably large plant screens. However, feasible and reliable polygenic genome engineering will need developments outside of the CRISPR editing machinery itself, including innovations in gRNA vector delivery for large cargos, and a broader conceptual shift toward population-level poly-gene editing leveraging large distributions of mutations for breeding, natural selection, or experimental evolution. Author ContributionsM.E.-A. conceived the project and secured funding. M.E.-A. and M.Es. designed the experimental strategy. M.Es. established the multiplex CRISPR and transformation pipelines in the laboratory, propagation through the T1 and T2 generations, and oversaw the first amplicon sequencing. Y.P. established the in-house iSeq amplicon sequencing protocol and contributed to cloning and genotyping pilots. M.Es supervised K.P. to construct cloning, bacterial transformations, plant growth, floral-dip transformations, and selection of T1 plants. E.K. contributed to early amplicon genotyping. J.K. propagated and sampled the T3 and J.K. and J.M. conducted the final amplicon sequencing panel. J.K. and M.E.-A. performed gene editing variant mapping, dataset quality control, and summarized results from published multiplex CRISPR studies. M.E.-A. modeled editing efficiency. J.K and M.E.-A. generated figures and wrote the first draft. All authors revised and improved the manuscript. J.K. and M.Es. contributed equally to this work and are designated as co-first authors.