Extended Data Fig. 10: Certain advantageous multi-KO tumour genotypes have limited accessibility of fitness trajectories.
From: Efficient and multiplexed somatic genome editing with Cas12a mice
a. Definitions of key fitness and epistasis metrics involving triple KO genotypes. b-c. Observed versus expected fitness values of all tumours containing arrays designed to create single, double or triple KOs, as measured in KT mice (lacking Cas12a and therefore with no actual KOs), where the expected values are calculated based on single KO effects only (b) or single KO and two-way epistasis effects (c). Medians with 95% confidence intervals are shown based on bootstrap resampling with 1,000 iterations. Dashed lines are identity lines. d-e. Fitness trajectories for all possible single, double and triple KO tumour genotypes involving Trp53, Keap1 and Mga (d), and Arid1a, Rb1 and Mga (e). Line colours indicate the successive fitness impact of inactivation of the indicated gene. 95% confidence intervals are shown for each genotype. Medians with 95% confidence intervals are shown based on bootstrap resampling with 1,000 iterations. Complete fitness and epistasis data for all arrays are shown in Supplementary Table 5.
