Extended Data Fig. 1: Sequencing errors can hinder the accurate identification of SNVs induced by prime editing.

a,b. Maps of lentiviral vectors used for the expression of prime editor (PEmax) (a) and pegRNAs (b). bpNLS, bipartite nuclear localization signal; MMLV-RT, human codon-optimized Moloney murine leukemia virus reverse transcriptase; ITR, inverted termainal repeat. (b) Locations of PCR primers used to deep-sequence a 229-bp region containing the RT template, PBS (primer binding site), pegRNA barcode, and unique molecular identifier (UMI) are shown. c, Proportion of sequencing reads containing substitution(s) in unedited PC-9 cells expressing NRCH-PEmax (left, unedited control) and in those ten days after transduction with the NRCH-exon20 library (right) for editing exon 20 of EGFR. d, Heatmap showing odds ratios and/or P-values of 558 (=186 ×3) SNVs generated by prime editing in exon 20 of EGFR ten days after the transuction of the NRCH-exon20 library. SNVs with P-values greater than 0.05 by the two-sided Fisher’s exact test are indicated in red; in these cases, odds ratios are not shown. SNVs with odds ratios lower than 3 are shown with a white background. The numbers at the bottom of each heatmap represent the location in the EGFR coding sequence. At each position, the nucleotide in the reference sequence is shown. e,f, Distribution of odds ratios (e) and P-values (f) of 558 SNVs in cells transduced with the NRCH-exon20 library. The dashed horizontal lines indicate the position at which the odds ratio = 3 (e) and the P-value = 0.05 (f). The P-value was calculated using a two-sided Fisher’s exact test. g, Distribution of observed SNV frequencies in PC-9 cells expresing PEmax ten days after transduction with NRCH-exon20. The number of SNVs n = 524 (Nonsignificant), n = 34 (Significant). Boxes represent the 25th, 50^th(median), and 75th percentiles, and whiskers show the 10th and 90th percentiles.