Fig. 5: Position-dependent editing efficiencies and indel profiles of PE2 and aPE2 across extended targets and SpCas9 variants.

a Base conversion efficiencies of PE2 and aPE2 at positions +7 to +14 at the HEK2 locus. b Indel at the same +7 to +14 positions as in (a). c Editing efficiencies of single-base conversions at positions +1 to +5 at the HEK2 locus. d, Editing efficiencies at positions +1, +2, +3, and +4 when co-edited with a + 5 G-to-A substitution at the same locus, representing conditional base conversions. e Base conversion efficiencies at the +5 position using SpCas9-PE2 (N₊₄G₊₅G₊₆ PAM) and SpG-PE2 (N₊₄G₊₅N₊₆ PAM), with or without AcrIIA5. f Base conversion efficiencies at the +6 position under the same conditions as in (e). g Proposed mechanism for AcrIIA5 enhance PE activity. In the PE system (left), the edited strand remains susceptible to re-cleavage by SpCas9 after the initial editing event. In contrast, in the aPE system (right), co-expression of AcrIIA5 partially inhibits SpCas9 activity, thereby reducing re-cleavage of the edited strand. h Correlation between DeepSpCas9 scores and aPE:PE fold change. i Editing efficiencies of PE and aPE systems at the selected loci with high or low DeepSpCas9 score. All experiments were conducted in HEK293T cells. The data in (a–f) were obtained from n = 3 independent biological replicates. Bars represent mean ± s.d. Source data are provided as a Source Data file.