Irysview software

Long-range scaffolding of the assembly contigs with optical mapping was achieved using the Irys optical mapping platform (BioNano Genomics). HMW DNA was isolated from 3-week-old leaf tissue of in vitro grown 60444 and TME3 cassava plants, embedded in thin agarose plugs according to the IrysPrep Kit and the plant tissue DNA isolation protocol (BioNano Genomics). DNA molecules were labeled using the NT.BspQI DNA-nicking enzyme by incorporation of fluorescent-dUTP nucleotides according to the IrysPrep nick-and-repair protocol (BioNano Genomics). DNA samples were aliquoted and quantitated using the Qubit Fluorimeter run in broad-range mode. The final samples were then loaded onto the IrysChips, linearized and visualized by the BioNano Irys molecule imaging instrument. Molecules > 150 kb were assembled de novo using the pairwise assembler provided by the IrysView software package (BioNano Genomics) with p value threshold of 10⁻⁹.

The raw SLT1.0 SMRT reads were corrected and assembled into sequence contigs using CANU with default parameters. The contigs were used for HERA assembly with the corrected SMRT reads. To identify sequence overlaps, all contigs and corrected reads were aligned all-against-all using Minimap2 [42 ] and BWA [43 (link)] with default parameters. The HERA-assembled super-contigs were combined with BioNano genome maps to generate hybrid maps using IrysView software (BioNano Genomics) with a minimum length of 150 kb. The resulting contigs were further clustered basing on the Hi-C data using 3D-DNA software [44 (link)] with the default parameters. Pilon [26 ] was used for further error correction with three rounds of polishing.

High-molecular weight (HMW) DNA was extracted from up to 5 g fresh leaf tissue using a modified Bionano Genomics protocol^{2 (link)}. Briefly, extracted HMW DNA was nicked with the enzyme Nt.BspQI (NEB, Beverly, MA, USA), fluorescently labeled, repaired, and stained overnight according to the Bionano Genomics nick-labeling protocol^{31 (link)}. KBS-Mac-74 nick-labeled DNA was run on a single flow cell on the Irys platform (Bionano Genomics, San Diego, CA, USA), for 90 cycles to generate 22.5 Gb raw data. The IrysView software (Bionano Genomics; version 2.5.1) was used to quality filter the raw data (>100 kb length, >2.75 signal/noise ratio) and molecules were assembled into contigs using the default “human genome” parameters. Resulting Bionano cmaps were compared against the different assemblies using Bionano RefAlign^{2 (link)}, and collapsed regions or artificial expansions were detected as structural variations using the structomeIndel.py script (https://github.com/RyanONeil/structome). Rates for FP and FN nicking sites were extracted from the .err files, and genome coverage was computed from the .xmap files.