Fig. 1

a In vitro CRISPR-Cas9 enrichment steps. High molecular weight nuclear DNA is extracted and crRNA probes designed. DNA 5′ ends are dephosphorylated to reduce ligation of sequencing adapters to non-target strands. Cas9 ribonucleoprotein particles (RNPs) bind at each side and cleave the region of interest (ROI). Double strand DNA is cleaved by Cas9 revealing blunt ends with ligatable 5′ phosphates. Cas9 remains bound to the protospacer adjacent motif (PAM)-distal end giving directionality for the strands towards the ROI. All DNA is dA-tailed, preparing the blunt ends for sequencing adapter ligation. ONT adapters are ligated to Cas9 cut sites which are both 3′ dA-tailed and 5′ phosphorylated. Library is cleaned to remove excess adapters using AMPure XP beads. Non-target molecules are not removed from the library. Library is added to the flow cell for sequencing. b Bioinformatics pipeline. Raw FAST5 reads are base called using Albacore2 (v2.3.4, ONT) and Guppy (v3.2.4, ONT) and converted to FASTQ. Reads get adapters trimmed using Porechop (v0.2.3 https://github.com/rrwick/Porechop), corrected using Canu (v1.7) [36] and aligned to the apple reference genome (‘Golden Delicious’ double haploid GDDH13v1.1) [34] using Minimap2 (v2.9) [46] to localize physically the ‘on-target’ and ‘off-target’ enriched regions. A de novo assembly is performed by Canu (v1.7) [36] using Albacore2 (v2.3.4, ONT) corrected reads and polished using Nanopolish (v0.11.1) [37]. Canu_corrected reads and the Canu_nanopolished assembly are used as inputs to run the final assembly performed by Flye (v2.5) [38]