Fig. 5: The engineering of TnpB guided by ProMEP enhances its editing efficiency in mammalian cells.

a Editing efficiency of TnpB variants harboring either top 10 beneficial or deleterious mutations at EMX1 site 1 is presented as the comparative fold change in InDel efficiency of these single-mutation variants relative to that of the WT TnpB. b Accuracy of the identification of both beneficial and detrimental single mutations in TnpB by ProMEP. c Editing efficiency of TnpB variants with triple mutations at EMX1 site 1. Data are presented as the fold change in InDel efficiency for TnpB variants with triple mutations relative to that of WT TnpB InDel efficiency. d The editing efficiency of WT TnpB and its quintuple mutants was assessed at three endogenous genomic loci in HEK293T cells. NC negative control. e Comparison of the editing efficiencies of TnpB and TnpB-AI-5.6 at 13 genomic loci in human HEK293T cells. NC negative control. f Distribution of deletions generated by the WT TnpB and TnpB-AI-5.6 in HEK293T cells at the AGBL1 site 1. The average efficiency of three biological replicates is symbolized by a single dot. g Schematic construct designs for ABEs derived from dTnpB and dTnpB-AI-5.6 with the WT TadA and ABE8e (TadA*), and miniature CBEs derived from dTnpB and dTnpB-AI-5.6 with the mutant APOBEC3A* (Y130F) and uracil glycosylase inhibitor (UGI). h A-to-G conversion efficiency in endogenous loci with ABEs derived from dTnpB and dTnpB-AI-5.6. i C-to-T conversion efficiency in endogenous loci with CBEs derived from dTnpB and dTnpB-AI-5.6. For a, c–e, h, i, data are means ± SD from three independent biological replicates.