Fig. 3: Characterization of PNLM-pcABE.
From: Protein-nucleic acid language model-assisted design of precise and compact adenine base editor
a The A-to-G base editing efficiency of ABE8e, ABE9 and PNLM-pcABE weas examined at 12 endogenous genomic loci containing multiple As in HEK293T cells. Heatmap reflects averaged data from three biological replicates. b The average A-to-G base editing efficiency of ABE8e, ABE9 and PNLM-pcABE at 12 endogenous genomic loci containing multiple As in Fig. 3a. Data are mean ± s.d. (n = 3 independent experiments) and p values were determined by a two-sided paired Wilcoxon rank-sum test. (ABE9 vs PNLM-pcABE: p = 0.0002 at A5, p = 0.0044 at A6, p = 0.0005 at A7) c The normalized ratio of highest/sub-optimal A-to-G base editing efficiency for ABE8e and PNLM-pcABE at the 12 target sites in Fig. 3a. Data are mean ± s.d. (n = 3 independent experiments). d The C-to-D (T/G/A) editing efficiency of ABE8e, ABE9 and PNLM-pcABE was examined at 12 endogenous genomic loci containing one C or multiple Cs in HEK293T cells. Heatmap reflects averaged data from three biological replicates. e The indel frequency formation of ABE8e, ABE9 and PNLM-pcABE at 21 endogenous genomic loci in Fig. 3a, d. Data are mean ± s.d. (n = 3 independent experiments) and p values were determined by a two-sided paired Wilcoxon rank-sum test (ABE8e vs ABE9: p = 0.0078; ABE8e vs PNLM-pcABE: p <0.0001; ABE9 vs PNLM-pcABE: p = 0.0012). Source data are provided as a Source data file.
